[Lynx-dev] Year 2038 problem in handling cookie date
I found a site sending a cookie whose gobble date is after January 2038. https://tabi-labo.com/ Lynx crashes when accessing the site with 32-bit system. I wrote an ad-hoc patch to prevent the crash. diff -rub orig/lynx2.8.9dev.16/src/parsdate.c lynx2.8.9dev.16/src/parsdate.c --- orig/lynx2.8.9dev.16/src/parsdate.c Sat Jan 05 11:00:30 2013 +++ lynx2.8.9dev.16/src/parsdate.c Sun Oct 28 17:34:05 2018 @@ -653,8 +653,16 @@ } Julian += tod; tod = Julian; -if (dst == DSTon || (dst == DSTmaybe && localtime()->tm_isdst)) +struct tm *tm; +tm = localtime(); +if ((tm != NULL) && (Julian >= 0)) { + if (dst == DSTon || (dst == DSTmaybe && tm->tm_isdst)) { Julian -= DST_OFFSET * 60 * 60; + } +} +else { + Julian = LONG_MAX; +} return Julian; } ___ Lynx-dev mailing list Lynx-dev@nongnu.org https://lists.nongnu.org/mailman/listinfo/lynx-dev
Re: [Lynx-dev] Small patches for lynx2.8.9dev.16
> The issue is that resize_term_WinPDC is compiled also by some compilers > that are not windows compilers. I would suggest a small modification > like the patch below. After this change everthing compiled and > worked flawlessly for me. Thanks again for your nice comment. These three functions were surrounded by #ifdef USE_MAXSCREEN_TOGGLE, and USE_MAXSCREEN_TOGGLE is defined only in Windows environment. int resize_term_WinPDC(int nlines, int ncols) void maxmizeWindowSize(void) void recoverWindowSize(void) But void start_curses(void) is a function for all environments, so I updated a patch. It is a replacement of lynx.patch_for_289dev16-9. From: Juan Manuel GuerreroSubject: Re: [Lynx-dev] Small patches for lynx2.8.9dev.16 Date: Mon, 28 Aug 2017 22:56:20 +0200 > Am 28.08.2017 18:00, schrieb Takeshi Hataguchi: >> Thanks for your useful comment. >> I agree with your comment regarding DOS like file systems. >> >>> bug produced when the coding is posix centric. It remains unclear >>> to me why stat is used instead of access. Something like: >>>access(dst, F_OK) != 0 >>> is faster than having to compute stat for a file on windows or dos >>> AFAIK. Do not get me wrong, for this application it will not really >>> matter. >> >> On the other hand, I'm not sure which function we should use for checking >> file existence. >> According to the following site, access() is used in "Compliant Solution >> (POSIX)", but _access_s() is used in "Compliant Solution (Windows)." There >> seems to be some reason not to use access() on Windows. >> https://www.securecoding.cert.org/confluence/display/c/FIO10-C.+Take+care+when+using+the+rename%28%29+function >> >> So I decided not to use access(), but it's not a strong opinion. >> >> Thanks, >> >> Takeshi Hataguchi > > Thanks for you replay. One small issue more. I have applied all the > patches proposed by you and compiled the source code using DJGPP. > That is a port of GCC to MSDOS/FreeDOS and I got the following error message: > > [snip] > gcc -DHAVE_CONFIG_H -DLOCALEDIR=\"/dev/env/DJDIR/share/locale\" -I. -I.. > -Ichrtrans -I.././../src/chrtrans -I.././.. -I.././../src > -I.././.././WWW/Library/Implementation-I/dev/env/WATT_ROOT/inc > -D_XOPEN_SOURCE=500 -D_POSIX_C_SOURCE=199506L -DDOSPATH -DNOUSERS > -I/dev/env/DJDIR/lib/../include -I/dev/env/DJDIR/include -g2 -O2 > -fgnu89-inline -march=i386 -mtune=i586 -save-temps -c .././../src/LYCurses.c > .././../src/LYCurses.c: In function 'start_curses': > .././../src/LYCurses.c:1549:2: warning: implicit declaration of function > 'resize_term_WinPDC'; did you mean 'resize_term'? > [-Wimplicit-function-declaration] > resize_term_WinPDC(scrsize_y, scrsize_x); > ^~ > resize_term > [snip] > Linking and creating Lynx executable > gcc -DHAVE_CONFIG_H -DLOCALEDIR=\"/dev/env/DJDIR/share/locale\" -I. -I.. > -Ichrtrans -I.././../src/chrtrans -I.././.. -I.././../src > -I.././.././WWW/Library/Implementation-I/dev/env/WATT_ROOT/inc > -D_XOPEN_SOURCE=500 -D_POSIX_C_SOURCE=199506L -DDOSPATH -DNOUSERS > -I/dev/env/DJDIR/lib/../include -I/dev/env/DJDIR/include -g2 -O2 > -fgnu89-inline -march=i386 -mtune=i586 -save-temps -L/dev/env/DJDIR/lib -o > lynx.exe LYebcdic.o LYClean.o LYShowInfo.o LYEdit.o LYStrings.o LYMail.o > HTAlert.o GridText.o LYGetFile.o LYMain.o LYMainLoop.o LYCurses.o > LYBookmark.o LYmktime.o LYUtils.o LYOptions.o LYReadCFG.o LYSearch.o > LYHistory.o LYForms.o LYPrint.o LYrcFile.o LYDownload.o LYNews.o LYKeymap.o > HTML.o HTFWriter.o HTInit.o DefaultStyle.o LYUpload.o LYLeaks.o LYexit.o > LYJump.o LYList.o LYCgi.o LYTraversal.o LYEditmap.o LYCharSets.o > LYCharUtils.o LYMap.o LYCookie.o LYStyle.o LYHash.o LYPrettySrc.o TRSTable.o > parsdate.o UCdomap.o UCAux.o UCAuto.o LYSession.o LYExtern.o LYLocal.o > .././WWW/L > ibrary/Implementation/libwww.a -lz -lbz2 -lpdcurses -L/dev/env/WATT_ROOT/lib > -lssl -lcrypto -lwatt -lz > LYCurses.o: In function `start_curses': > l:/_/lynx2.8.9dev.16/_/src/.././../src/LYCurses.c:1549: undefined reference > to `resize_term_WinPDC' > collect2.exe: error: ld returned 1 exit status > make.exe[1]: *** [Makefile:111: lynx.exe] Error 1 > make.exe[1]: Leaving directory 'l:/_/lynx2.8.9dev.16/_/src' > make.exe: *** [Makefile:221: all] Error 2 > > The issue is that resize_term_WinPDC is compiled also by some compilers > that are not windows compilers. I would suggest a small modification > like the patch below. After this change everthing compiled and > worked flawlessly for me. > > Regards, > Juan M. Guerrero > > > > > diff -aprNU5 lynx2.8.9dev.16.orig/src/LYCurses.c > lynx2.8.9dev.16/src/LYCurses.c > --- lynx2.8.9dev.16.orig/src/LYCurses.c 2017-08-27 01:12:12 + > +++ lynx2.8.9dev.16/src/LYCurses.c2017-08-28 22:25:58 + > @@ -974,10 +974,11 @@ int saved_winpos_y2 = 0; > > #ifdef USE_MAXSCREEN_TOGGLE > static HWND currentWindowHandle = NULL; > static char
Re: [Lynx-dev] Small patches for lynx2.8.9dev.16
Thanks for your useful comment. I agree with your comment regarding DOS like file systems. > bug produced when the coding is posix centric. It remains unclear > to me why stat is used instead of access. Something like: > access(dst, F_OK) != 0 > is faster than having to compute stat for a file on windows or dos > AFAIK. Do not get me wrong, for this application it will not really > matter. On the other hand, I'm not sure which function we should use for checking file existence. According to the following site, access() is used in "Compliant Solution (POSIX)", but _access_s() is used in "Compliant Solution (Windows)." There seems to be some reason not to use access() on Windows. https://www.securecoding.cert.org/confluence/display/c/FIO10-C.+Take+care+when+using+the+rename%28%29+function So I decided not to use access(), but it's not a strong opinion. Thanks, Takeshi Hataguchi From: Juan Manuel GuerreroSubject: Re: [Lynx-dev] Small patches for lynx2.8.9dev.16 Date: Sun, 27 Aug 2017 22:19:55 +0200 > lynx.patch_for_289dev16-3 discribes a really serious issue but not > only for windows but also for all other DOS like file systems too, > thus I would suggest something like the small fix shown below. It > shows a modification of the original patch. It is is a classical > bug produced when the coding is posix centric. It remains unclear > to me why stat is used instead of access. Something like: > access(dst, F_OK) != 0 > is faster than having to compute stat for a file on windows or dos > AFAIK. Do not get me wrong, for this application it will not really > matter. > > Regards, > Juan M. Guerrero > > > > > > diff -aprNU5 lynx2.8.9dev.16.orig/src/LYUtils.c lynx2.8.9dev.16/src/LYUtils.c > --- lynx2.8.9dev.16.orig/src/LYUtils.c2017-08-27 00:17:50 + > +++ lynx2.8.9dev.16/src/LYUtils.c 2017-08-27 10:19:32 + > @@ -7102,13 +7102,13 @@ void LYAddHtmlSep0(char *path) > * Rename a file > */ > int LYRenameFile(char *src, > char *dst) > { > -#ifdef _WINDOWS > +#ifdef DOSPATH > /* > - * If dest_file exists prior to calling rename(), rename() will fail on > Windows platforms. > + * If dest_file exists prior to calling rename(), rename() will fail on > Windows and DOS platforms. > * > https://www.securecoding.cert.org/confluence/display/c/FIO10-C.+Take+care+when+using+the+rename%28%29+function > */ > struct stat st; > if (stat(dst, ) == 0) { > unlink(dst); > > ___ > Lynx-dev mailing list > Lynx-dev@nongnu.org > https://lists.nongnu.org/mailman/listinfo/lynx-dev ___ Lynx-dev mailing list Lynx-dev@nongnu.org https://lists.nongnu.org/mailman/listinfo/lynx-dev
[Lynx-dev] Small patches for lynx2.8.9dev.16
I wrote some small patches for lynx2.8.9dev.16. If some of them conflicts with your policy, please discard them. Thanks. - Terminate when pressing 'd'(download) on no action button. lynx.patch_for_289dev16 NoActionButtonTest.html - Incompatibility in move() between POSIX and Windows. ref: https://www.securecoding.cert.org/confluence/display/c/FIO10-C.+Take+care+when+using+the+rename%28%29+function lynx.patch_for_289dev16-3 - Can't show long title in the title bar on Windows environment. lynx.patch_for_289dev16-6 - Wrapping line behavior problem with PDCurses. ref: lynx-dev discussions "Subject: Wrapping line behavior" lynx.patch_for_289dev16-7 - Resizing terminal problem with Windows + PDCurses. This problem occurs only with some combinations of source and destination screen size. For example: 80x25 -> 90x20 # including some code clean up lynx.patch_for_289dev16-9 Takeshi Hataguchi lynx.patch_for_289dev16 Description: Binary data foo lynx.patch_for_289dev16-3 Description: Binary data lynx.patch_for_289dev16-6 Description: Binary data lynx.patch_for_289dev16-7 Description: Binary data lynx.patch_for_289dev16-9 Description: Binary data ___ Lynx-dev mailing list Lynx-dev@nongnu.org https://lists.nongnu.org/mailman/listinfo/lynx-dev
Re: [Lynx-dev] Wrapping line behavior
>> > as well as knowing what X resources you are using for kterm (as well as >> > the $TERM value -- which should be "kterm", not "xterm", since the latter >> > uses several features _not_ in kterm's repertoire). >> >> How can I show X resources I'm using for kterm? >> I don't use .Xresources and command line is "kterm -km euc -geometry >> 80x25+0+0&" > > then I'm still puzzled. I ran your config.status/config.cache against > 2.8.8dev.9 and 2.8.8dev.10 trees on my Ubuntu 14.04 machine, eliminated > environment variables (LYNX_CFG, TERMINFO, TERMINFO_DIRS) which might > affect the behavior, and still do not see lynx failing to wrap the lines. Thanks. But it's very very weird. By the way, I tried -dump option, and got the result that only 288dev10 fails to wrap line. % ./lynx -dump -cfg=/dev/null WrapLineTest.html So I think the terminal and the setting regarding to terminal aren't the cause of this problem. -- Takeshi Hataguchi E-mail: patak...@users.sourceforge.jp Onside PRE .10203040.50...6070 8090100...110...120...130...140 ...150...160...170...180...190...200 __ Iutside PRE .10203040.50...60708 090100...110...120...130...140...150...1 60...170...180...190...200 __ IFRAME IFRAME: [1]/10203040.50...60... .708090100...110...120...130... 140...150...160...170...180...190...200 References 1. file://localhost/10203040.50...60708090100...110...120...130...140...150...160...170...180...190...200 Onside PRE .10203040.50...60708090100...110...120...130...140...150...160...170...180...190...200 __ Iutside PRE .10203040.50...60708090100...110...120...130...140...150...160...170...180...190...200 __ IFRAME IFRAME: [1]/10203040.50...60708090100...110...120...130...140...150...160...170...180...190...200 References 1. file://localhost/10203040.50...60708090100...110...120...130...140...150...160...170...180...190...200 ___ Lynx-dev mailing list Lynx-dev@nongnu.org https://lists.nongnu.org/mailman/listinfo/lynx-dev
Re: [Lynx-dev] Wrapping line behavior
> So... it would be nice to have a copy of these files from your configure > and build: > > a) config.log > > b) config.status > > c) config.cache Could you find attached files? > as well as knowing what X resources you are using for kterm (as well as > the $TERM value -- which should be "kterm", not "xterm", since the latter > uses several features _not_ in kterm's repertoire). How can I show X resources I'm using for kterm? I don't use .Xresources and command line is "kterm -km euc -geometry 80x25+0+0&" The $TERM was "kterm" in the kterm. -- Takeshi Hataguchi E-mail: patak...@users.sourceforge.jp config_logs.tar.gz Description: Binary data ___ Lynx-dev mailing list Lynx-dev@nongnu.org https://lists.nongnu.org/mailman/listinfo/lynx-dev
Re: [Lynx-dev] Wrapping line behavior
From: Thomas DickeySubject: Re: [Lynx-dev] Wrapping line behavior Date: Thu, 2 Feb 2017 20:37:55 -0500 > c) Your suggested change modifies (undoes) an older chunk of code dating >from early 2012. Here's my change note for that: > > REV:1.219 GridText.c 2012/02/05 18:55:54 > tom You are right. That is as I intended. I wrote "Wrapping line behavior seems to have been changed since 2.8.8dev.10." Not since 2.8.9dev.10 but 2.8.8dev.10. I have updated my lynx binary recently for the first time since 2011. -- Takeshi Hataguchi E-mail: patak...@users.sourceforge.jp ___ Lynx-dev mailing list Lynx-dev@nongnu.org https://lists.nongnu.org/mailman/listinfo/lynx-dev
Re: [Lynx-dev] Wrapping line behavior
>> 2017/01/19 08:33 ... X dej: >>> When using the equal key (=), the URL after "Action:" of this page >>> titled "Information about the current document" can be, when too long >>> to fit, wrapped respecting the margins. This is unfortunate because it >>> is not easy to select that URL with a mouse or with other standard >>> copy-paste behaviors. Lynx has a feature to copy document's/link's URL to clipboard. I think it's better to use it in this situation. I'm setting as below. - Environment variable export RL_CLCOPY_CMD='/usr/bin/xsel -ib' - lynx.cfg KEYMAP:C:TO_CLIPBOARD -- Takeshi Hataguchi E-mail: patak...@users.sourceforge.jp >> >> "Standard?" > > Tmux and Screen have copy-paste included, use the shortcuts '[' and > ']' (square brackets). > > They are also fooled by these spaces at linewraps. > > I just tested that what is printed by command echo > averyvery[...]longline can be mouse selected by one double click, > inside tmux. > >> In Lynx I never use the mouse--never need to--one reason why >> I like it. > > When your lynx is showing a web page sounding like the included video > is interesting, or a web page malfunctionning because it need > javascript for what you want to do, how do you usually react ? > > I assure you that I use numbers to select links[1]. > > X dej > > [1] the small problem is when the first link of the currently shown > page is a text field: I need to get of that text field first before > typing a link number. > > ___ > Lynx-dev mailing list > Lynx-dev@nongnu.org > https://lists.nongnu.org/mailman/listinfo/lynx-dev ___ Lynx-dev mailing list Lynx-dev@nongnu.org https://lists.nongnu.org/mailman/listinfo/lynx-dev
Re: [Lynx-dev] Wrapping line behavior
>> For the unconditional wrapping - I suppose it's possible that lynx is >> confused about the width of the terminal. If it were only kterm (and >> not also the Windows port), I'd check if you have disabled the auto-wrap >> feature (which kterm, being based on xterm, I think would apply). > > It doesn't depend on the terminal, I got the same results with kterm, xterm > and gnome-terminal. > I think lynx isn't confused about the width of the terminal because long line > consist of plural words is wrapped as expected. > The problem occurs with long line consist of single word. I wrote a patch against it. It works fine for me. I tested following environments. - 2.8.8dev10 + slangon Ubuntu - 2.8.8dev10 + ncursesw on Ubuntu (No degrade check) - 2.8.9dev11 + slangon Ubuntu - 2.8.9dev11 + ncursesw on Ubuntu (No degrade check) - 2.8.9dev11 + PDCurses on Windows -- Takeshi Hataguchi E-mail: patak...@users.sourceforge.jp lynx.patch_for_288dev10 Description: Binary data ___ Lynx-dev mailing list Lynx-dev@nongnu.org https://lists.nongnu.org/mailman/listinfo/lynx-dev
Re: [Lynx-dev] Wrapping line behavior
> For the unconditional wrapping - I suppose it's possible that lynx is > confused about the width of the terminal. If it were only kterm (and > not also the Windows port), I'd check if you have disabled the auto-wrap > feature (which kterm, being based on xterm, I think would apply). It doesn't depend on the terminal, I got the same results with kterm, xterm and gnome-terminal. I think lynx isn't confused about the width of the terminal because long line consist of plural words is wrapped as expected. The problem occurs with long line consist of single word. >> By the way, I can see this problem with Windows' binary using PDCurses. >> I used following binary, and screen shot is attached. >> http://invisible-island.net/datafiles/release/lynx-cs-setup.exe >> >> With Windows' binary, I found a little bit different result, that is the >> long text with link is wrapped only when it is selected. >> Viewing WrapLineTest2.html and moving cursor, "some text" is vanished. >> This seems to happen only with Windows + PDCurses. > > I'll check that on my Windows machine (thanks). Thanks, -- Takeshi Hataguchi E-mail: patak...@users.sourceforge.jp ___ Lynx-dev mailing list Lynx-dev@nongnu.org https://lists.nongnu.org/mailman/listinfo/lynx-dev
Re: [Lynx-dev] Wrapping line behavior
>> > I tried 2.8.8dev.10 with ncursesw just now, and then got result as same as >> > 2.8.8dev.9 (former expected result). >> > I confirmed this problem with slang on Ubuntu and PDCurses2.5 on Windows. >> >> I just tried with slang (and 2.8.8dev.10) and see the older behavior: >> >> Lynx Version 2.8.9dev.12 (24 Nov 2016) >> libwww-FM 2.14, SSL-MM 1.4.1, OpenSSL 1.0.1t, s-lang 2.2.4 >> Built on linux-gnu (Jan 1 2017 21:35:11). Thanks, but I can't understand why your result differs to mine. I also use s-lang 2.2.4 and my configure script is below. ./configure --with-screen=slang By the way, I can see this problem with Windows' binary using PDCurses. I used following binary, and screen shot is attached. http://invisible-island.net/datafiles/release/lynx-cs-setup.exe With Windows' binary, I found a little bit different result, that is the long text with link is wrapped only when it is selected. Viewing WrapLineTest2.html and moving cursor, "some text" is vanished. This seems to happen only with Windows + PDCurses. -- Takeshi Hataguchi E-mail: patak...@users.sourceforge.jp Title: Wrap Line Test Onside PRE .10203040.50...60708090100...110...120...130...140...150...160...170...180...190...200 some text Iutside PRE .10203040.50...60708090100...110...120...130...140...150...160...170...180...190...200 some text IFRAME some text ___ Lynx-dev mailing list Lynx-dev@nongnu.org https://lists.nongnu.org/mailman/listinfo/lynx-dev
Re: [Lynx-dev] Wrapping line behavior
Thanks for your comment. It seems to depend on screen library. I tried 2.8.8dev.10 with ncursesw just now, and then got result as same as 2.8.8dev.9 (former expected result). I confirmed this problem with slang on Ubuntu and PDCurses2.5 on Windows. By the way, the version I reported is correct. It's 2.8.8dev.{9,10}, not 2.8.9dev.{9,10}. -- Takeshi Hataguchi E-mail: patak...@users.sourceforge.jp From: Thorsten GlaserSubject: Re: [Lynx-dev] Wrapping line behavior Date: Sun, 15 Jan 2017 15:18:54 + (UTC) > patak...@users.sourceforge.jp dixit: > >>Attached files are the test file and screen shots of 2.8.8dev.9 and >>2.8.8dev.10. >>I tried with slang. > > Hmm, in Debian (2.8.9dev.11), it looks the same as in MirBSD > (2.8.8dev.16); both use ncursesw. Both look like your “before” > so did you mean 2.8.9dev.{9,10}? > > bye, > //mirabilos > -- > Stéphane, I actually don’t block Googlemail, they’re just too utterly > stupid to successfully deliver to me (or anyone else using Greylisting > and not whitelisting their ranges). Same for a few other providers such > as Hotmail. Some spammers (Yahoo) I do block. ___ Lynx-dev mailing list Lynx-dev@nongnu.org https://lists.nongnu.org/mailman/listinfo/lynx-dev
[Lynx-dev] Wrapping line behavior
Wrapping line behavior seems to have been changed since 2.8.8dev.10. I think old behavior is preferable because all characters are shown. Attached files are the test file and screen shots of 2.8.8dev.9 and 2.8.8dev.10. I tried with slang. -- Takeshi Hataguchi E-mail: patak...@users.sourceforge.jp Title: Wrap Line Test Onside PRE .10203040.50...60708090100...110...120...130...140...150...160...170...180...190...200 Iutside PRE .10203040.50...60708090100...110...120...130...140...150...160...170...180...190...200 IFRAME ___ Lynx-dev mailing list Lynx-dev@nongnu.org https://lists.nongnu.org/mailman/listinfo/lynx-dev
[Lynx-dev] LYE_PASTE doesn't work properly on Windows
LYE_PASTE doesn't work properly on Windows. Attached file is an ad-hoc patch to solve the problem. I'm sorry that I'm not sure it's appropriate way. -- Takeshi Hataguchi E-mail: patak...@users.sourceforge.jp lynx.patch_for_289dev11-2 Description: Binary data ___ Lynx-dev mailing list Lynx-dev@nongnu.org https://lists.nongnu.org/mailman/listinfo/lynx-dev
Re: [Lynx-dev] Unexpected behavior downloading Lynx's binary
> Then you write that content served with a "Content-Encoding: gzip" > header "should" be decompressed by lynx, which, to me, implied that > it wasn't happening in some cases. But that wasn't what the > original poster was talking about. I'm sorry to confuse you and the delay in my reply. I wanted to report 2 unexpected behaviors in downloading file. 1. Lynx saves the transmission stream data (gzipped data) when the content is served with a "Content-Encoding: gzip" header. 2. Lynx adds the file extention "gzip" in some cases and doesn't add any file extention in other cases in the situation same as 1. So I think Thorsten wrote the same problem as 1. -- Takeshi Hataguchi E-mail: patak...@users.sourceforge.jp From: L A WalshSubject: Re: [Lynx-dev] Unexpected behavior downloading Lynx's binary Date: Sun, 01 Jan 2017 17:13:25 -0800 > Thorsten Glaser wrote: >> L A Walsh dixit: >> >>> The "Content-Encoding" gzip is a feature of many web-server software >>> packages. Your client would automatically uncompress it before >>> storing it. This is independent of any file format. >> That’s precisely what I said, thank you for repeating it. > --- > Did you mean in this message?: > > Thorsten Glaser wrote: >> No. >> >> What he means (I think — I occasionally experience a similar problem) >> is that content served with a “Content-Encoding: gzip” header should >> be decompressed by lynx before being saved by the (D)ownload process >> (or passed to a DOWNLOADER utility). > --- > Since I wasn't clear what you were saying. > You said you experience a similar problem, which confused > me, as I didn't see what the problem was. > > Then you write that content served with a "Content-Encoding: gzip" > header "should" be decompressed by lynx, which, to me, implied that > it wasn't happening in some cases. But that wasn't what the > original poster was talking about. It seemed they were confused > about the use of "Content...gzip" in the header and the fact that the files > were bzip2 or "zoo" -- and that they were not > gzipped files. > > I was saying that the two are completely separate - and, unlike > what you wrote, I didn't have or see any problem with such. I > still don't know what problem you experience -- but I didn't see it being the > same as the poster problem which was just being > "confused". If that was your problem, as well, then I understand > why I might be confused about what you meant! :-) > > Or were you expressing a bit of sarcasm in your thanking me for > repeating your point? > > cheers! > -linda > > p.s. I only joined the list shortly before I posted my question > about how to get coloring to work, so if you posted anything > before that, I didn't see it... > > > ___ > Lynx-dev mailing list > Lynx-dev@nongnu.org > https://lists.nongnu.org/mailman/listinfo/lynx-dev ___ Lynx-dev mailing list Lynx-dev@nongnu.org https://lists.nongnu.org/mailman/listinfo/lynx-dev
[Lynx-dev] patch for treatment Japanese characters
I wrote a patch to correct treatment Japanese characters. It seems to be broken by integration (2.8.8dev.9) and modification against warning (2.8.8dev.6). -- Takeshi Hataguchi E-mail: patak...@users.sourceforge.jp lynx.patch_for_289dev11 Description: Binary data ___ Lynx-dev mailing list Lynx-dev@nongnu.org https://lists.nongnu.org/mailman/listinfo/lynx-dev
[Lynx-dev] Unexpected behavior downloading Lynx's binary
I found an unexpected behavior in downloading Lynx's binary by Lynx. http://invisible-mirror.net/archives/lynx/tarballs/lynx2.8.9dev.11.tar.bz2 --> Downloaded file: lynx2.8.9dev.11.tar.bz2 The file is gzipped. (mismatch between content (gzipped) and file extention (bz2)) http://invisible-mirror.net/archives/lynx/tarballs/lynx2.8.9dev.11.zip --> Downloaded file: lynx2.8.9dev.11.zip.gz The file is gzipped but I expect not gzipped file, that is lynx2.8.9dev.11.zip. I tried it using 2.8.9dev.11 on Ubuntu 14.04. Configure option includes "--with-bzlib --with-zlib" Preferred encoding: All I found a discussion below. https://lists.nongnu.org/archive/html/lynx-dev/2016-08/msg3.html Is this also a server problem? According to the trace log, the server sent "Content-Encoding: gzip" in the response header in both cases. I think the client should store uncompressed files in these cases. ___ Lynx-dev mailing list Lynx-dev@nongnu.org https://lists.nongnu.org/mailman/listinfo/lynx-dev
[Lynx-dev] trouble with showing 1 line page at the bottom
The current Lynx has a trouble to show 1 line page at the bottom of the document. The page indicator is also wrong. For example, when display_lines is 23 and a document has 24 lines, it shows only 23 lines but can't go to the next page. The page indicator shows (p1 of 1). I wrote an ad hoc patch and attached. -- Takeshi Hataguchi E-mail: [EMAIL PROTECTED] lynx.patch_for_286rel4 Description: Binary data ___ Lynx-dev mailing list Lynx-dev@nongnu.org http://lists.nongnu.org/mailman/listinfo/lynx-dev
[Lynx-dev] strange behavior opening localfile with #
I've got a report about Lynx's strange behavior opening localfile with # on Windows. These are the behaviors which I found, putting a file, which name is foo#bar, on the current directory and d:/tmp. 1. It SHOULD be opend and actually it is. (expected) lynx -cfg=lynx.cfg -lss=lynx.lss foo#bar 2. It SHOULD be opend but actually it isn't. (unexpected) lynx -cfg=lynx.cfg -lss=lynx.lss ./foo#bar lynx -cfg=lynx.cfg -lss=lynx.lss d:/tmp/foo#bar lynx -cfg=lynx.cfg -lss=lynx.lss file://localhost/d:/tmp/foo%23bar 3. It SHOULDN'T be opend and actually it isn't. (expected) lynx -cfg=lynx.cfg -lss=lynx.lss ./foo%23bar lynx -cfg=lynx.cfg -lss=lynx.lss d:/tmp/foo%23bar lynx -cfg=lynx.cfg -lss=lynx.lss file://localhost/d:/tmp/foo#bar 4. It SHOULDN'T be opend but actually it is. (unexpected) lynx -cfg=lynx.cfg -lss=lynx.lss ./foo%2523bar lynx -cfg=lynx.cfg -lss=lynx.lss d:/tmp/foo%2523bar lynx -cfg=lynx.cfg -lss=lynx.lss file://localhost/d:/tmp/foo%2523bar I've tested 2.8.6rel.1 for Windows got from this URL. http://pervalidus.50webs.org/cygwin/lynx/ -- Takeshi Hataguchi E-mail: [EMAIL PROTECTED] ___ Lynx-dev mailing list Lynx-dev@nongnu.org http://lists.nongnu.org/mailman/listinfo/lynx-dev
[Lynx-dev] hyperlink's color with color-style
I found a trouble with hyperlink's color with color-style. hyperlinkColorTest.html has 11 hyperlinks. 1. Open it, image's color is as the same as normal text. http://lynx-win32-pata.sourceforge.jp/tmp/hyperlinkColorTest_fig1.png # not expected. 2. And press down-arrow 10 times, then #link, 2, 5 and 8's color change to as the same as normal text. http://lynx-win32-pata.sourceforge.jp/tmp/hyperlinkColorTest_fig2.png # not expected. 3. And press Ctrl+L, then 2, 5 and 8's color become as expected but #link's color is still wrong. http://lynx-win32-pata.sourceforge.jp/tmp/hyperlinkColorTest_fig3.png I got the same unexpected results with 2.8.6rel.2 on Debian stable with ncurses, color-style and D.C.S. is euc-jp 2.8.6pre.3 on Debian stable with ncurses, color-style and D.C.S. is euc-jp 2.8.6rel.1 on Windows with PDCurses, color-style and D.C.S. is Shift_JIS # get from http://pervalidus.50webs.org/cygwin/lynx/ command line: ./lynx -cfg=foo.cfg -lss=/dev/null hyperlinkColorTest.html On the other hand, I got an expected results with 2.8.6pre.3 on Debian stable with s-lang and without color-style. command line: ./lynx -cfg=foo.cfg hyperlinkColorTest.html So I think this is a problem concerned with color-sytle. -- Takeshi Hataguchi E-mail: [EMAIL PROTECTED] hyperlinkColorTest.html Description: Binary data foo.cfg Description: Binary data ___ Lynx-dev mailing list Lynx-dev@nongnu.org http://lists.nongnu.org/mailman/listinfo/lynx-dev
Re: [Lynx-dev] nested-table in lynx2.8.6pre.2
On Fri, 15 Sep 2006, [EMAIL PROTECTED] wrote: Also, can lynx-devers suggest some URLs of good examples of tables where nested-table logic improves the presentation? Thanks. http://weather.asahi.com/pinpoint/kanto/tokyo.html Tokyo's wether information page in Japanese. I can get better result with a large screen. Tried http://weather.asahi.com/pinpoint/kinki/nara.html;. :) Hmmm. Maybe there is something wrong with my configuration of Lynx. I don't see any appreciable difference between nested-table toggled on or off at run time. The rendering as a whole is very bad either way, to the point that the table is practically useless. Cells are displayed vertically instead of horizontally. (Is it my screen? I access Lynx via a shell account with PuTTY on WindowsXP as the terminal emulator; only 80 characters wide.) Please try wider screen. I tried with 120x35. It's not perfect but so so. The cells are aligned like stairs but we can understand the weather information when netsted-table is on. -- Takeshi Hataguchi E-mail: [EMAIL PROTECTED] ___ Lynx-dev mailing list Lynx-dev@nongnu.org http://lists.nongnu.org/mailman/listinfo/lynx-dev
Re: [Lynx-dev] nested-table in lynx2.8.6pre.2
On Tue, 12 Sep 2006, [EMAIL PROTECTED] wrote: On Tue, 12 Sep 2006, Henry Nelson wrote: On Mon, Sep 11, 2006 at 05:01:13PM -0400, Thomas Dickey wrote: 2001-08-15 (2.8.5dev.2) * work around defect in move_anchors_in_region() and related logic by changing default for nested-tables to FALSE when Lynx is not configured for color-style. The problem is that when an anchor is shifted right by nested-table logic, if it has a BR near the beginning of a table cell and it happens to be split across a line, its size will not be adjusted properly (report by Hataguchi Takeshi) -TD There are two questions to answer: (a) whether we should set the default to enable nested tables, given that for most pages it is an improvement, or disable it by default to lessen confusion. Is nested-table logic only supposed to work with, i.e., depend on, color-style, which in turn assumes screen=ncurses, or is it that it has only been tested with color-style and may work in other configurations? no - it works either way, but has been tested more with color-style. The context for the bug noted above was that there was some logic difference between the two styles that happened to make it work better. I tested 2.8.6pre.3 with ncurses and slang on Debian. color-style is enable for only the former binary. I couldn't find any differences between them with my example file except default setting. Presently I configure lynx with --with-screen=slang and --enable- nested-tables, but I've always wondered if that is the right thing to be doing since I don't set a configuration option for color-style, and always assumed it was off by default. (b) whether the run-time toggling of the feature works properly. With Lynx 2.8.6dev.18 (I'll do my best to try to upgrade to lynx2.8.6 pre.3 this weekend.), --with-screen=slang and --enable-nested-tables, I get a status-line message that toggling is ON/OFF. 2.8.6pre.3 also shows the message and the setting makes effect with above two binaries. # That is, I saw a problem when it's on with both binaries. # Columns is 80. Assuming I can get lynx2.8.6pre.3 installed soon, what test should we do to see if run-time toggling of the feature works properly. Also, can lynx-devers suggest some URLs of good examples of tables where nested-table logic improves the presentation? Thanks. http://weather.asahi.com/pinpoint/kanto/tokyo.html Tokyo's wether information page in Japanese. I can get better result with a large screen. I find that it improves google's search page (but that it's not good for some email admin webpages). ymmv -- Thomas E. Dickey http://invisible-island.net ftp://invisible-island.net ___ Lynx-dev mailing list Lynx-dev@nongnu.org http://lists.nongnu.org/mailman/listinfo/lynx-dev -- Takeshi Hataguchi E-mail: [EMAIL PROTECTED] ___ Lynx-dev mailing list Lynx-dev@nongnu.org http://lists.nongnu.org/mailman/listinfo/lynx-dev
Re: [Lynx-dev] nested-table in lynx2.8.6pre.2
On Sun, 10 Sep 2006, [EMAIL PROTECTED] wrote: On Mon, 11 Sep 2006, [EMAIL PROTECTED] wrote: On Sat, 9 Sep 2006, [EMAIL PROTECTED] wrote: On Sat, Sep 09, 2006 at 06:00:25PM +0900, [EMAIL PROTECTED] wrote: It seems that nested-table is on by default in lynx2.8.6pre.2 with configure option --nable-nested-tables, though it's off by default until lynx2.8.6dev.18. I think it's better off by default because the bug, I reported on 02 May 2003, doesnt seem to be fixed yet. There is the compile-time default and the run-time default. iirc, they have almost the same effect. Does your case work well enough if we turn off the run-time default? It does work well at the start time, but doesn't after turning on by pressing '~', even if we turn off the run-time default. oh (that could be something I can improve). Do you have an example that I can study? I attached a sample html file to this mail. screenshot: http://lynx-win32-pata.sourceforge.jp/tmp/nestedTables.png At the moment, I added a lynx.cfg entry to allow overriding without the command-line option. -- Thomas E. Dickey http://invisible-island.net ftp://invisible-island.net -- Takeshi Hataguchi E-mail: [EMAIL PROTECTED] nestedTables.html Description: Binary data ___ Lynx-dev mailing list Lynx-dev@nongnu.org http://lists.nongnu.org/mailman/listinfo/lynx-dev
[Lynx-dev] nested-table in lynx2.8.6pre.2
It seems that nested-table is on by default in lynx2.8.6pre.2 with configure option --nable-nested-tables, though it's off by default until lynx2.8.6dev.18. I think it's better off by default because the bug, I reported on 02 May 2003, doesnt seem to be fixed yet. http://lists.gnu.org/archive/html/lynx-dev/2003-05/msg00017.html -- Takeshi Hataguchi E-mail: [EMAIL PROTECTED] ___ Lynx-dev mailing list Lynx-dev@nongnu.org http://lists.nongnu.org/mailman/listinfo/lynx-dev
Re: [Lynx-dev] UTF-8 problem
On Thu, 07 Sep 2006, [EMAIL PROTECTED] wrote: Hi all, I'm recently migrating from ja_JP.eucJP to ja_JP.UTF-8 and might find a problem. On kterm with ja_JP.eucJP, lynx can display a page of both ISO-2022-JP or euc-jp (or even UTF-8) if a source file setts a correct charset=ISO-2022-JP or charset=euc-jp. But on uxterm with ja_JP.UTF-8, it seems lynx fails to display a page of ISO-2022-JP but can dispaly correctly a page of euc-jp. (precisely, lynx doesn't display a page of euc-jp completely correctly but it can dispaly japanese characters well.) It must fail because there is no mechanism to convert iso-2022-jp to utf-8 in current Lynx code. see CHANGELOG | 2005-10-17 (2.8.6dev.14) (snip) | * extend experimental option --enable-japanese-utf8, allowing lynx to convert | EUC-JP and Shift_JIS strings to UTF-8 -TH and my mail. http://lists.gnu.org/archive/html/lynx-dev/2005-07/msg00018.html I think adding iso-2022-jp to utf-8 support is too large task for me. I'm hesitating to do it because I don't have enough time. # I think that first, I have to add the definition in UCdomap.[ch], # then check the incluences on existing codes and then introduce the # conversion. I don't know well but if utf-8 terminal emulator will become standard in the near future, Lynx might be an outdated program for Japanese. # Though we can use Lynx with TTYConv utility. Please try http://www1.pm.tokushima-u.ac.jp/~kohda/ then select TeX (the first link so only pressing Enter key) and then select XeTeX (again the first link). The first page is in English so no problem. the TeX is in ISO-2022-JP and lynx displays series of symbols the XeTeX is in euc-jp and lynx displays Japanese characters. I'm afraid this could be a simple problem of settings or my HTML files might be wrong but anyway please enlighten me. Regards, 2006-9-7(Thu) -- Debian Developer Debian JP Developer - much more I18N of Debian Atsuhito Kohda kohda AT debian.org Department of Math., Univ. of Tokushima ___ Lynx-dev mailing list Lynx-dev@nongnu.org http://lists.nongnu.org/mailman/listinfo/lynx-dev -- Takeshi Hataguchi E-mail: [EMAIL PROTECTED] ___ Lynx-dev mailing list Lynx-dev@nongnu.org http://lists.nongnu.org/mailman/listinfo/lynx-dev
Re: Compiling MingW lynx under Cygwin (was Re: [Lynx-dev] text browsing)
On Sun, 30 Jul 2006, [EMAIL PROTECTED] wrote: On Sat, 29 Jul 2006 [EMAIL PROTECTED] wrote: May I ask you the concrete steps to compile Lynx under Cygwin with -mno-cygwin? Do you mean running configure make on Cygwin bash? OK. I went back through my notes and patches and did some more testing. Here is a step-by-step guide to building a statically compiled lynx under Cygwin which will not be dependent on the cygwin dll (i.e. a MingW binary). There were a lot of patches, so this is long, but with all the details it should avoid the need for multiple followup questions. Let me know if anything isn't clear or if it doesn't work as claimed on your machine. Doug Thanks a lot. I tried it and finally successed to build with some efforts. I wrote some comments and questions below. Compiling lynx for MingW under Cygwin This post will try to outline how I built Windows lynx for MingW, using the -mno-cygwin directive under Cygwin. I haven't looked carefully at each of the steps to see whether or not it is important for a successful build. Please remember that I am not a programmer, so not all of the changes that I made may be appropriate. Before beginning to try this, I would suggest reading Mumit Khan's paper on how to use -mno-cygwin. Unfortunately, his web page about MingW appears to have disappeared from the net. A copy of the HOWTO has been saved and is available from DJ Delorie's web site at http://www.delorie.com/howto/cygwin/mno-cygwin-howto.html;. Note that using -mno-cygwin turns off scanning of /usr/include and /usr/local/include in favor of /usr/include/mingw and /usr/local/include/mingw. This assumes that you have an install of Cygwin, including all the usual utilities, and with installation of the mingw modules. When I built my mingw programs, I frequently did so with $prefix=/mingw, so some includes and libraries are installed in /mingw/include and /mingw/lib rather than in /usr/include/mingw, /usr/local/include/mingw, etc., so appropriate paths have to be specified when compiling to be sure that the mingw files are found, rather than the cygwin files. The general plan is as follows: 1. Compile and install MingW PDCurses 2.8 2. Compile and install MingW OpenSSL 0.9.8 stable branch 3. Compile and install MingW iconv 4. Compile and install MingW gettext I want to use precompiled libraries if it's OK. If I should to use patched libraries, please tell me which libraries sould be patched. 5. Patch Cygwin autoconf to comply with Tom Dickey's patches. Note that standard autoconf will fail with the lynx configure.in script. 6. Patch aclocal.m4 for DOS/Windows compatibility 7. Run autoconf to regenerate configure 8. Run configure with the appropriate options to generate the makefiles 9. If desired, compile a lynx icon into a coff file with windres and modify the src/makefile to include the icon file in the lynx binary 10. Run make to generate lynx and make install-full to install 11. Optionally compress the lynx binary with upx The details: 1. Compile and install MingW PDCurses 2.8 I used precompiled library: pdcurses-2.6.0-2003.07.21-1.exe I got from http://www.mingw.org/download.shtml. 2. Compile and install MingW OpenSSL 0.9.8 stable branch I applied this patch before compiling: I tried GnuTLS instead of OpenSSL. I used precompiled library: gnutls_1_0_17_win32_mingw_libs.zip I got from ftp://ftp.gnutls.org/pub/gnutls/win32. 3. Compile and install MingW iconv I used libiconv 1.8. I applied this patch: I used precompiled library: libiconv-1.9.1.bin.woe32.zip I got from http://sourceforge.net/project/showfiles.php?group_id=25167 4. Compile and install MingW gettext I used gettext-0.11.5. I applied this patch: I used precompiled library: gettext-0.11.5-2003.02.01-1.exe I got from http://www.mingw.org/download.shtml. # I found two different source archives of gettext-0.11.5 from # http://www.gnu.org/software/gettext/ # http://www.mingw.org/download.shtml # # I failed to apply your patch to the former one. # I successed to apply it the latter one but haven't successed to built. 5. Patch Cygwin autoconf to comply with Tom Dickey's patches. I used cygwin autoconf 2.59-2. First I applied the cygwin patch that comes with the source code. I then applied a patch combining the DOS pathname patches from the DJGPP port and a modification of the patch on Tom Dickey's web page for autoconf 2.57 (http://invisible-island.net/autoconf/autoconf.html;). This is the patch: [snip] Then run configure, then make, then make install. No problem. 6. Patch aclocal.m4 for DOS/Windows compatibility Finally to lynx, itself. Apply the patch I recently submitte to the lynx-dev list for DOS path compatibility. No problem. 7. Run autoconf to regenerate configure Next, regenerate the configure file by running autoconf. Unfortunately, autoconf 2.59 is no longer
Re: [Lynx-dev] text browsing
On Fri, 28 Jul 2006, [EMAIL PROTECTED] wrote: The MingW version of lynx which I have been porting (but not yet distributing) is compiled with NLS and seems to work fine. I linked it with gettext-0.11.5 which I ported to MingW last year. I generally compile MingW programs under Cygwin with -mno-cygwin and don't have any experience with the native MingW MSYS environment. May I ask you the concrete steps to compile Lynx under Cygwin with -mno-cygwin? Do you mean running configure make on Cygwin bash? I've tried these but I'm not satisfied yet. (1) using makelynx.bat on command prompt (2) configure make under MSYS envrironment (1) It rebuilds all whenever I run the batch file because it doesn't check the timestamp like make command does. (2) I can build it with slang but can't with pdcurses. I want to use pdcurses because I can change screen size only with pdcurses. So I'm very curious to know how you compile Lynx. Thanks in advance. -- Takeshi Hataguchi E-mail: [EMAIL PROTECTED] ___ Lynx-dev mailing list Lynx-dev@nongnu.org http://lists.nongnu.org/mailman/listinfo/lynx-dev
Re: [Lynx-dev] text browsing
On Tue, 25 Jul 2006, [EMAIL PROTECTED] wrote: [EMAIL PROTECTED] dixit: - change the compiler from Borland's bcc32 to MingW's gcc Why? It depends on SSL library. I think it's easier to compile with GnuTLS using the MingW's compiler than using Borland's one. - change the SSL library from OpenSSL to GnuTls NO! That'd be BAD! It should be done for the license problem. - write the English project page You could at least add an (international) behind the link on the page. Thanks. I updated the page. bye, //mirabile -- I believe no one can invent an algorithm. One just happens to hit upon it when God enlightens him. Or only God invents algorithms, we merely copy them. If you don't believe in God, just consider God as Nature if you won't deny existence.-- Coywolf Qi Hunt -- Takeshi Hataguchi E-mail: [EMAIL PROTECTED] ___ Lynx-dev mailing list Lynx-dev@nongnu.org http://lists.nongnu.org/mailman/listinfo/lynx-dev
Re: [Lynx-dev] Problem with SSL certificates in Lynx
X509_NAME_oneline doesn't seem to be compatible with GnuTLS and OpenSSL. Lynx doesn't work well with GnuTLS's format. [snip] Debian stable uses gnutls by default... Aki Tuomi That was already reported to Debian Bug tracking system 2 years ago. There also seems to be patches and a personal package but it isn't closed yet. http://bugs.debian.org/cgi-bin/bugreport.cgi?bug=231609 -- Takeshi Hataguchi E-mail: [EMAIL PROTECTED] ___ Lynx-dev mailing list Lynx-dev@nongnu.org http://lists.nongnu.org/mailman/listinfo/lynx-dev
Re: [Lynx-dev] Problem with SSL certificates in Lynx
On Mon, 26 Jun 2006, [EMAIL PROTECTED] wrote: I tried looking thru the mailing list archives and could not find anything useful for the following issue. When connecting to a site with a proper certificate (not self-signed, using a proper CA, listed properly in /etc/ssl/certs) i get error SSL error:Can't find common name in certificate-Continue? It seems that Lynx built with GnuTLS always shows this improper message when connecting any SSL sites. X509_NAME_oneline doesn't seem to be compatible with GnuTLS and OpenSSL. Lynx doesn't work well with GnuTLS's format. OpenSSL: /C=JP/ST=Tokyo/(snip)/CN=... GnuTLS : C=, ST=Tokyo,(snip), CN=... I tested OpenSSL/0.9.6l on Windows and gnutls11 (1.0.16-13.2) on Debian 3.1. This error does not exhibit itself in other browsers, nor when testing with openssl -CApath /etc/ssl/certs -connect site:443 I've set SSL_CERT_DIR=/etc/ssl/certs I am now wondering why it won't work when it's done properly. This is most annoying since I am using a commercially signed certificate. If you need more information please don't hesitate to ask. Just to ignore the message, to set SSL Prompting as force yes-response in option menu. Aki Tuomi ___ Lynx-dev mailing list Lynx-dev@nongnu.org http://lists.nongnu.org/mailman/listinfo/lynx-dev -- Takeshi Hataguchi E-mail: [EMAIL PROTECTED] ___ Lynx-dev mailing list Lynx-dev@nongnu.org http://lists.nongnu.org/mailman/listinfo/lynx-dev
Re: [Lynx-dev] Lynx with SSL support for WinXP installation needed
To Whom It May Concern: I'm trying to get a version of Lynx that can read SSL pages on the web on a WinXP box. I am not technical, so I'd like an easy install method if that is possible. Download Lynx286dev17TH.exe from https://sourceforge.jp/projects/lynx-win32-pata/files/ This is an easy install and setup package, I believe. Any help in this matter would be greatly appreciated. Sean Leong Online Product Marketing Manager Verified Identity Pass 1270 Avenue of the Americas, Suite 508 New York, New York 10020 Office: +1.212.332.6365 Mobile: +1.973.632.2307 Fax: +1.212.332.6308 -- Takeshi Hataguchi E-mail: [EMAIL PROTECTED] ___ Lynx-dev mailing list Lynx-dev@nongnu.org http://lists.nongnu.org/mailman/listinfo/lynx-dev
[Lynx-dev] compile error under Cygwin (was: lynx2.8.6dev.18)
I tried to compile 2.8.6dev.18 under Cygwin environment but failed. First, I failed to compile HTTCP.c. It's improved after applying a patch which is attached to this mail. Second, I failed to link with these messages. | Linking and creating Lynx executable | gcc -DWIN_EX -D_WINDOWS_NSL -I/usr/local/include -U_XOPEN_SOURCE -D_XOPEN_SOURCE=600 -U_POSIX_C_SOURCE -D_POSIX_C_SOURCE=199506L -DHAVE_CONFIG_H -DLOCALEDIR=\/usr/local/share/locale\ -I. -I.. -Ichrtrans -I./chrtrans -I.. -I../src -I../WWW/Library/Implementation -O2 -o lynx.exe LYClean.o LYShowInfo.o LYEdit.o LYStrings.o LYMail.o HTAlert.o GridText.o LYGetFile.o LYMain.o LYMainLoop.o LYCurses.o LYBookmark.o LYUtils.o LYOptions.o LYReadCFG.o LYSearch.o LYHistory.o LYForms.o LYPrint.o LYrcFile.o LYDownload.o LYNews.o LYKeymap.o HTML.o HTFWriter.o HTInit.o DefaultStyle.o LYUpload.o LYLeaks.o LYexit.o LYJump.o LYList.o LYCgi.o LYTraversal.o LYEditmap.o LYCharSets.o LYCharUtils.o LYMap.o LYCookie.o LYStyle.o LYHash.o LYPrettySrc.o TRSTable.o UCdomap.o UCAux.o UCAuto.o LYLocal.o ../WWW/Library/Implementation/libwww.a -lslang -L/usr/local/lib -luser32 | ../WWW/Library/Implementation/libwww.a(HTTCP.o):HTTCP.c:(.text+0x5f2): undefined reference to `_WSASetLastError' | Info: resolving _h_errno by linking to __imp__h_errno (auto-import) | collect2: ld returned 1 exit status | make[1]: *** [lynx.exe] Error 1 | make[1]: Leaving directory `/cygdrive/d/home/patakuti/tmp/cygwin/dev18/lynx2-8-6/src' | make: *** [all] Error 2 Why undefined reference to `_WSASetLastError' ? How can I fix it? configure option: --with-screen=slang % uname -a CYGWIN_NT-5.1 myPC 1.5.19(0.150/4/2) 2006-01-20 13:28 i686 Cygwin -- Takeshi Hataguchi E-mail: [EMAIL PROTECTED] lynx.patch_for_286dev18 Description: Binary data ___ Lynx-dev mailing list Lynx-dev@nongnu.org http://lists.nongnu.org/mailman/listinfo/lynx-dev
[Lynx-dev] patch for cygwin
I wrote a patch for compiling under cygwin environment. - modify not to check d_ino when __CYGWIN__ is defined because it doesn't seem to be supported by Cygwin now. - modify to use fill_rehostent to avoid segfault in _fork_func, which is used Win32 environment (includes Cygwin). - modify not to use WSASetLastError when __CYGWIN__ is defined to avoid link error. # This is ad hoc change. I test it with 2.8.6dev.17. -- Takeshi Hataguchi E-mail: [EMAIL PROTECTED] lynx.patch_for_286dev17-2 Description: Binary data ___ Lynx-dev mailing list Lynx-dev@nongnu.org http://lists.nongnu.org/mailman/listinfo/lynx-dev
[Lynx-dev] Lynx settings page is vanished and RULE:MAP behavior
Lynx settings page: http://www.hippo.ru/%7Ehvv/lynxcfg_toc.html, which is linked from Lynx Help Page, seems to be vanished. I heard that a person copied the pages in http://www.geocities.jp/lynx_cfg/lynxcfg_toc.html. # He is not an original author. lynx_help/Lynx_users_guide.html and lynx_help/lynx_help_main.html should be modified. By the way, I found RULE:Redirect http://www.hippo.ru/%7Ehvv/* http://www.geocities.jp/lynx_cfg/* works as expected, but RULE:MAP http://www.hippo.ru/%7Ehvv/* http://www.geocities.jp/lynx_cfg/* doesn't work as expected. In the latter case, Lynx seems to send below http request to www.geocities.jp. GET /lynx_cfg/lynxcfg_toc.html HTTP/1.0\r Host: www.hippo.ru\r Accept: text/html, text/plain, text/sgml, */*;q=0.01\r Accept-Encoding: gzip, compress, bzip2\r Accept-Language: ja\r User-Agent: Lynx/2.8.6dev.17TH libwww-FM/2.14FM SSL-MM/1.4.1 OpenSSL/0.9.6l\r I think Host: www.hippo.ru isn't proper here, but Host: www.geocities.jp is. Is this a bug or a restriction? -- Takeshi Hataguchi E-mail: [EMAIL PROTECTED] ___ Lynx-dev mailing list Lynx-dev@nongnu.org http://lists.nongnu.org/mailman/listinfo/lynx-dev
Re: [Lynx-dev] Japanese character vanished in right side edge
On Sat, 21 Jan 2006, [EMAIL PROTECTED] wrote: On Sat, Jan 21, 2006 at 06:05:19PM +0900, [EMAIL PROTECTED] wrote: I received a report that a Japanese character was vanished in the right side edge. I attached a sample file for this problem and put screen shots below. what character set does the sample file use? (there's no header). euc-jp and I set D.C.S. to euc-jp. http://lynx-win32-pata.sourceforge.jp/tmp/rightSideEdgeTest286dev16.png Double width 0 is vanished in the right side edge. This is unexpected. Lynx: 2.8.6dev.16 with no patch Is this with pdcurses? slang on sid. There may be no problem with pdcurses on Windows. -- Takeshi Hataguchi E-mail: [EMAIL PROTECTED] ___ Lynx-dev mailing list Lynx-dev@nongnu.org http://lists.nongnu.org/mailman/listinfo/lynx-dev
Re: [Lynx-dev] Japanese character vanished in right side edge
On Sun, 22 Jan 2006, I wrote: On Sat, 21 Jan 2006, [EMAIL PROTECTED] wrote: On Sat, Jan 21, 2006 at 06:05:19PM +0900, [EMAIL PROTECTED] wrote: I received a report that a Japanese character was vanished in the right side edge. I attached a sample file for this problem and put screen shots below. what character set does the sample file use? (there's no header). euc-jp and I set D.C.S. to euc-jp. http://lynx-win32-pata.sourceforge.jp/tmp/rightSideEdgeTest286dev16.png Double width 0 is vanished in the right side edge. This is unexpected. Lynx: 2.8.6dev.16 with no patch Is this with pdcurses? slang on sid. Sorry. I wanted to write slang on sarge. There may be no problem with pdcurses on Windows. -- Takeshi Hataguchi E-mail: [EMAIL PROTECTED] -- Takeshi Hataguchi E-mail: [EMAIL PROTECTED] ___ Lynx-dev mailing list Lynx-dev@nongnu.org http://lists.nongnu.org/mailman/listinfo/lynx-dev
[Lynx-dev] curius behaviors navigating a page internally
I found curius behaviors of Lynx2.8.6dev.16. case1: 1. start Lynx with command line: navigatePageTest.html#l50, then Lynx shows line 50. % lynx navigatePageTest.html#l50 2. press tab key, then cursor is moved to line 100. 3. press ^R key, then screen is moved to line 50. This is unexpected. It's expected that screen isn't moved. case2: 1. start Lynx with command line: navigatePageTest.html#l50, then Lynx shows line 50. % lynx navigatePageTest.html#l50 2. press tab key, then cursor is moved to line 100. 3. press Enter key on link to line 0, then screen is moved to line 0. 4. press Left Arrow(PREV_DOC), then screen is moved to line 50. This is unexpected. It's expected that screen is moved to line 100. -- Takeshi Hataguchi E-mail: [EMAIL PROTECTED] navigatePageTest.html Description: Binary data ___ Lynx-dev mailing list Lynx-dev@nongnu.org http://lists.nongnu.org/mailman/listinfo/lynx-dev
Re: [Lynx-dev] security problem in EXTERNAL on Windows
On Sat, 26 Nov 2005, I wrote: I found a security problem in some versions of lynx, includes 2.8.5rel.1 and 2.8.6dev.15. There may be some risks when executing external programs against malicious URL by EXTERNAL mechanism on Windows platform. I wrote a patch for 2.8.6dev.15 and tested on Windows XP. I did only on Windows XP but the behavior is different on Win95/98 and WinNT/2000/XP. So I'm happy if someone test on above platforms. I updated my patch because Lynx passed the path, in which some characters are encoded, to the external program when the URL is file scheme with the last patch. For example, file://localhost/c:/foo%20bar.txt is encoded to c:/foo^%20bar.txt and it's passed to the external program. It's safe but many programs can't recognize it as a proper local path. They expect c:/foo bar.txt. # It seems that slash(/) instead of backslash(\) doesn't matter here. On the other hand it's ok for other scheme except file, so I made it have other transaction for file scheme than others. And I have to write that we should to use start command or batch file as EXTERNAL program. Or special characters in the parameter are passed to a program as escaped form. # Special characters (, ^, %) are escaped by ^. # I hadn't know that the handling against special character are # difference by normal commands and start command or batch file. Please note this patch is for original 2.8.6dev.15. Don't apply both this and former one. -- Takeshi Hataguchi E-mail: [EMAIL PROTECTED] lynx.patch_for_286dev15-2 Description: Binary data ___ Lynx-dev mailing list Lynx-dev@nongnu.org http://lists.nongnu.org/mailman/listinfo/lynx-dev
[Lynx-dev] Option Menu with 2.8.6dev.15 on Windows
I have a trouble with showing Option Menu with 2.8.6dev.15 on Windows, which I compiled with PDCurses 2.5. This is a screen shot with trouble. http://lynx-win32-pata.sourceforge.jp/tmp/286dev15_OptionScreen.png For example, Case insensitive is showed double. -- Takeshi Hataguchi E-mail: [EMAIL PROTECTED] ___ Lynx-dev mailing list Lynx-dev@nongnu.org http://lists.nongnu.org/mailman/listinfo/lynx-dev
Re: [Lynx-dev] lynx display buffer questions
On Tue, 24 May 2005 I wrote: # I wrote details before. # http://lists.gnu.org/archive/html/lynx-dev/2004-02/msg00201.html Sorry; I had forgotten. What is the status of your iconv patch that you mention in that message? This patch has been already merged into Lynx's code since 2.8.6dev.4 (--enable-japanese-utf8). But this makes it possible only to convert utf-8 strings to euc-jp or Shift_JIS. Reverse conversion isn't implemented yet because I thought it was needed seriously. I wrote an additional patch concerned with this. http://www.t3.rim.or.jp/~patakuti/tmp/lynx/lynx.patch_for_286dev13 This patch makes lynx to be able to convert euc-jp and Shift_JIS strings to utf-8. But unfortunately not to covert iso-2022-jp to utf-8. # need more hack to do it. Anyway I think this patch gives us(Japanese) chance to enjoy WWW more. For example, using msn search, Wikipedia's search on euc-jp screen and viewing euc-jp pages on UTF-8 screen. So to view euc-jp pages in utf-8 screen, some work (writing code) is needed. -- Takeshi Hataguchi E-mail: [EMAIL PROTECTED] -- Takeshi Hataguchi E-mail: [EMAIL PROTECTED] ___ Lynx-dev mailing list Lynx-dev@nongnu.org http://lists.nongnu.org/mailman/listinfo/lynx-dev
Re: [Lynx-dev] lynx display buffer questions
# I wrote details before. # http://lists.gnu.org/archive/html/lynx-dev/2004-02/msg00201.html Sorry; I had forgotten. What is the status of your iconv patch that you mention in that message? This patch has been already merged into Lynx's code since 2.8.6dev.4 (--enable-japanese-utf8). But this makes it possible only to convert utf-8 strings to euc-jp or Shift_JIS. Reverse conversion isn't implemented yet because I thought it was needed seriously. So to view euc-jp pages in utf-8 screen, some work (writing code) is needed. -- Takeshi Hataguchi E-mail: [EMAIL PROTECTED] ___ Lynx-dev mailing list Lynx-dev@nongnu.org http://lists.nongnu.org/mailman/listinfo/lynx-dev
Re: [Lynx-dev] lynx display buffer questions
I was able to read UTF-8 pages and see Cyrillic, Japanese, Hebrew, Arabic and Korean all on the same page (http://czyborra.com/) in Lynx 2.8.6dev.7. ** Pretty fantastic. ** Good. _BUT_, now I can't view pages written in euc-jp !!! Now Lynx can't translate text from euc-jp to utf-8, so you can't view such pages with UTF-8 setting. # I wrote details before. # http://lists.gnu.org/archive/html/lynx-dev/2004-02/msg00201.html Also, I couldn't get UTF-8 to work in a screen session. It is SO frustrating. Try -U command line option, or utf8 on on command in screen session. # press ^A:utf8 on onEnter -- Takeshi Hataguchi E-mail: [EMAIL PROTECTED] ___ Lynx-dev mailing list Lynx-dev@nongnu.org http://lists.nongnu.org/mailman/listinfo/lynx-dev
[Lynx-dev] unescape '\' in mailcap's command
Since 2.8.6dev.5, Lynx seems to unescape '\' twice in command line which is indicated in mailcap file. For example, we have to write image/jpeg; C\:APPIVIEWI_VIEW32.EXE %s to call C\:\APP\IVIEW\I_VIEW32.EXE. Until 2.8.6dev.4, we can write image/jpeg; C\:\\APP\\IVIEW\\I_VIEW32.EXE %s to call the same command. Is this a change with any reason? I like previous behavior. -- Takeshi Hataguchi E-mail: [EMAIL PROTECTED] ___ Lynx-dev mailing list Lynx-dev@nongnu.org http://lists.nongnu.org/mailman/listinfo/lynx-dev
Re: [Lynx-dev] lynx display buffer questions
On Sun, 15 May 2005 16:14:29, [EMAIL PROTECTED] wrote: On Thu, May 12, 2005 at 11:04:27PM +0900, [EMAIL PROTECTED] wrote: available. The best I can do is the Japanese MS Gothic, which gives me English, Japanese and some Cyrillic and Greek. Do you mean that MS Gothic is enough for your use? It's not really enough. I do read enough French, and a tad of German and Spanish, that I'd like to be able to see glyphs for those languages. Until a fixed-width Unicode font becomes available for Windows, that's not going to happen. I see your problem. I found BDF UM+, which is a fixed-width Unicode font and its TTF file can be get from http://www.kaoriya.net/. I can use this font with PuTTY and see UTF-8 documents. *I* think looking is soso, but moving cursor is bad (highlighted text is duplicated). I also found unifont, which is an another fixed-width Unicode font. It's distributed in HEX file but I could change it to TTF file using hex2bdf and bdf2ttf utilities. And it also can be used with PuTTY. # fontsize 12 seems to be fine. http://czyborra.com/unifont/unifont.hex.gz http://czyborra.com/unifont/hex2bdf http://www.kaoriya.net/dist/bdf2ttf-1.0.002.tar.bz2 I tested with Lynx : 2.8.4rel.1 Server: Debian (Woody) Client: Windows2000 and PuTTY Release 0.54-jp20040423 Lynx's version may be old. -- Takeshi Hataguchi E-mail: [EMAIL PROTECTED] ___ Lynx-dev mailing list Lynx-dev@nongnu.org http://lists.nongnu.org/mailman/listinfo/lynx-dev
Re: [Lynx-dev] lynx display buffer questions
Just my $0.02, but as a Windows98/XP terminal emulator user, the main problem is that, to my knowledge, there is no fixed-width unicode font available. The best I can do is the Japanese MS Gothic, which gives me English, Japanese and some Cyrillic and Greek. Do you mean that MS Gothic is enough for your use? If yes, you can use it on PuTTY with jppatch(*). UTF-8 is also available on it. (*) http://hp.vector.co.jp/authors/VA024651/#PuTTYkj_top Thus, even if I have Lynx display the raw UTF8, the terminal emulator has no way to present it. Setting D.C.S to UNICODE(UTF-8) is useful for pages encoded by UTF-8. Of course, you need to use terminal emulator which is available unicode font. -- Takeshi Hataguchi E-mail: [EMAIL PROTECTED] ___ Lynx-dev mailing list Lynx-dev@nongnu.org http://lists.nongnu.org/mailman/listinfo/lynx-dev
Re: [Lynx-dev] lynx2.8.6dev.6
Thorsten Glaser wrote: [EMAIL PROTECTED] dixit: - most japanese pages work in kterm (note that my lynx is still English, and neither locale nor gettext/nls support exists in the operating environment) most = these in EUC-JP - some work if I set both lynx and kterm to display character set Shift-JIS (lynx seems to not convert between them) some work? Do you mean some work and some doesn't work? some = these in SJIS ;-) sorry for being unclear Really? I think there is no problem in viewing SJIS pages with Shift_JIS setting. Please confirm your setting and show me a concrete example if you still have troubles. - utf-8 pages don't work in kterm (no conversion) Please try --enable-japanese-utf8 configure option which is introduced from 2.8.6dev.4. Oh, how nice? It makes us be able to view the Japanese pages encoded UTF-8 properly with kterm and other similar terminal emulators. *looks* if test $use_ja_utf8 != no ; then AC_DEFINE(EXP_JAPANESEUTF8_SUPPORT) AC_CHECK_LIB(iconv,libiconv_open) fi Not nice. This library is not in the base installation, but lynx is. I cannot introduce a dependency on a port into the base system. Last time I checked, GNU libiconv was under LGPL. If someone knows of a free iconv library that works with lynx, please let me know. I believe that glibc's iconv and GNU libiconv does work. -- Takeshi Hataguchi E-mail: [EMAIL PROTECTED] ___ Lynx-dev mailing list [EMAIL PROTECTED] http://lists.nongnu.org/mailman/listinfo/lynx-dev
Re: [Lynx-dev] lynx2.8.6dev.6
Thorsten Glaser wrote: [EMAIL PROTECTED] dixit: - some work if I set both lynx and kterm to display character set Shift-JIS (lynx seems to not convert between them) some work? Do you mean some work and some doesn't work? some = these in SJIS ;-) sorry for being unclear Really? I think there is no problem in viewing SJIS pages with Shift_JIS setting. Yes, that's what I said. For EUC pages, it works, for SJIS pages, I have to switch kterm. Which do you mean switch from other terminal emulator to kterm, or switch kterm setting? If you mean the latter, no, you don't have to change kterm setting. You can use the same setting for SJIS pages as for EUC pages. That is using euc-jp setting for both of Lynx and kterm. # Using Shift_JIS setting for both is also OK. -- Takeshi Hataguchi E-mail: [EMAIL PROTECTED] ___ Lynx-dev mailing list [EMAIL PROTECTED] http://lists.nongnu.org/mailman/listinfo/lynx-dev
Re: [Lynx-dev] progress/status
I'm about 2/3 through the dev.6 changes, and am working on the slice that I started a couple of months ago - making the options menu display properly for Japanese. There are some repainting problems with the link-traversal, but the layout is improved. See ftp://invisible-island.net/temp/lynx286devf5.png Does this screenshot mean that dev.6 will be able to show euc-jp (or Shift_JIS) page correctly on UTF-8 terminal? If yes, Great! Thanks for your effort! -- Takeshi Hataguchi E-mail: [EMAIL PROTECTED] ___ Lynx-dev mailing list [EMAIL PROTECTED] http://lists.nongnu.org/mailman/listinfo/lynx-dev
Re: [Lynx-dev] Japanese (JIS, EUC, Shift-JIS), uxterm
On Fri, 16 Jul 2004, I wrote: Henry, please add this line to your lynx.cfg, then you should never see the extra charset meta in the downloaded file. PREPEND_BASE_TO_SOURCE:FALSE ouch! I wanted to write PREPEND_CHARSET_TO_SOURCE:FALSE. -- Takeshi Hataguchi E-mail: [EMAIL PROTECTED] ___ Lynx-dev mailing list [EMAIL PROTECTED] http://lists.nongnu.org/mailman/listinfo/lynx-dev