[Bug-wget] Support of non-linux OS's going down the drain?
As of wget-1.13, you seem to prefer GNUTLS over openssl. Whatever the reasoning behind it might be, the configuration now just FAILs if the target system only has OpenSSL and no GNU TLS. OK, I'll redefine that fail to it disables ssl when there is openssl. That is bad. Why? GNU TLS /might/ be more safe than OpenSSL in some aspects, but is is for sure not available on (older) versions of AIX and/or HP-UX. It is already quite a bit of work to get OpenSSL and OpenSSH to be rather actual/recent on those boxes, but you can simply forget getting gnutls to be available on those. The dependency chain is a straight hell. That said, even with (in 1.13 undocumented, but documented in 1.13.1 - thanks) --with-ssl=openssl, wget simply fails on all my HP-UX boxes. With HP-UX 11.00 and HP C-ANSI-C it doesn't even *compile* anymore! $ ./configure --prefix=/pro --disable-nls --with-ssl=openssl --without-libiconv-prefix --without-libintl-prefix --without-libgnutls-prefix : $ make : cc -DHAVE_CONFIG_H -I. -I../src -I/pro/local/include -I/usr/local/include -Ae -O2 +Onolimit +Z -z -I/pro/local/include -I/usr/local/include -I/usr/include/X11R6 -I/usr/local/X11R6/include -I/usr/contrib/X11R6/include -c -o c-ctype.o c-ctype.c source='cloexec.c' object='cloexec.o' libtool=no \ DEPDIR=.deps depmode=hp /bin/sh ../build-aux/depcomp \ cc -DHAVE_CONFIG_H -I. -I../src -I/pro/local/include -I/usr/local/include -Ae -O2 +Onolimit +Z -z -I/pro/local/include -I/usr/local/include -I/usr/include/X11R6 -I/usr/local/X11R6/include -I/usr/contrib/X11R6/include -c -o cloexec.o cloexec.c cpp: ./, line 4: warning 2014: Illegal ^A, ^B, ^C, or ^D in source. Changed to space. cpp: ./, line 7: warning 2014: Illegal ^A, ^B, ^C, or ^D in source. Changed to space. cpp: ./, line 13: warning 2014: Illegal ^A, ^B, ^C, or ^D in source. Changed to space. cpp: ./, line 21: warning 2014: Illegal ^A, ^B, ^C, or ^D in source. Changed to space. cc: , line 2: error 1000: Unexpected symbol: �. cc: , line 4: error 1000: Unexpected symbol: . cc: , line 4: error 1000: Unexpected symbol: . cc: , line 4: error 1000: Unexpected symbol: . cc: , line 4: error 1000: Unexpected symbol: . cc: , line 4: error 1000: Unexpected symbol: �. cc: , line 4: error 1000: Unexpected symbol: �. cc: , line 4: error 1000: Unexpected symbol: . cc: , line 4: error 1000: Unexpected symbol: . cc: , line 4: error 1000: Unexpected symbol: |. cc: , line 4: error 1000: Unexpected symbol: . cc: , line 4: error 1000: Unexpected symbol: . cc: , line 4: error 1000: Unexpected symbol: `. cc: , line 6: error 1000: Unexpected symbol: . cc: , line 7: error 1000: Unexpected symbol: p. cc: , line 13: error 1000: Unexpected symbol: �. cc: , line 16: error 1000: Unexpected symbol: . cc: , line 18: error 1000: Unexpected symbol: . cc: , line 20: error 1000: Unexpected symbol: . cc: , line 21: error 1000: Unexpected symbol: $float. cc: panic 2017: Cannot recover from earlier errors, terminating. make[4]: *** [cloexec.o] Error 1 make[4]: Leaving directory `/pro/3gl/GNU/wget-1.13.1/lib' make[3]: *** [all-recursive] Error 1 make[3]: Leaving directory `/pro/3gl/GNU/wget-1.13.1/lib' make[2]: *** [all] Error 2 make[2]: Leaving directory `/pro/3gl/GNU/wget-1.13.1/lib' make[1]: *** [all-recursive] Error 1 make[1]: Leaving directory `/pro/3gl/GNU/wget-1.13.1' make: *** [all] Error 2 Exit 2 -- H.Merijn Brand http://tux.nl Perl Monger http://amsterdam.pm.org/ using 5.00307 through 5.14 and porting perl5.15.x on HP-UX 10.20, 11.00, 11.11, 11.23 and 11.31, OpenSuSE 10.1, 11.0 .. 11.4 and AIX 5.2 and 5.3. http://mirrors.develooper.com/hpux/ http://www.test-smoke.org/ http://qa.perl.org http://www.goldmark.org/jeff/stupid-disclaimers/
Re: [Bug-wget] Support of non-linux OS's going down the drain?
can you please H.Merijn Brand h.m.br...@xs4all.nl writes: That is bad. Why? GNU TLS /might/ be more safe than OpenSSL in some aspects, but is is for sure not available on (older) versions of AIX and/or HP-UX. It is already quite a bit of work to get OpenSSL and OpenSSH to be rather actual/recent on those boxes, but you can simply forget getting gnutls to be available on those. The dependency chain is a straight hell. but in that case, a --with-ssl=openssl will fix this problem, as you did. With HP-UX 11.00 and HP C-ANSI-C it doesn't even *compile* anymore! $ ./configure --prefix=/pro --disable-nls --with-ssl=openssl --without-libiconv-prefix --without-libintl-prefix --without-libgnutls-prefix : $ make : cc -DHAVE_CONFIG_H -I. -I../src -I/pro/local/include -I/usr/local/include -Ae -O2 +Onolimit +Z -z -I/pro/local/include -I/usr/local/include -I/usr/include/X11R6 -I/usr/local/X11R6/include -I/usr/contrib/X11R6/include -c -o c-ctype.o c-ctype.c source='cloexec.c' object='cloexec.o' libtool=no \ DEPDIR=.deps depmode=hp /bin/sh ../build-aux/depcomp \ cc -DHAVE_CONFIG_H -I. -I../src -I/pro/local/include -I/usr/local/include -Ae -O2 +Onolimit +Z -z -I/pro/local/include -I/usr/local/include -I/usr/include/X11R6 -I/usr/local/X11R6/include -I/usr/contrib/X11R6/include -c -o cloexec.o cloexec.c cpp: ./, line 4: warning 2014: Illegal ^A, ^B, ^C, or ^D in source. Changed to space. cpp: ./, line 7: warning 2014: Illegal ^A, ^B, ^C, or ^D in source. Changed to space. cpp: ./, line 13: warning 2014: Illegal ^A, ^B, ^C, or ^D in source. Changed to space. cpp: ./, line 21: warning 2014: Illegal ^A, ^B, ^C, or ^D in source. Changed to space. can you please send me your cloexec.c file (or any other file causing it)? My cloexec.c doesn't have such fancy characters, and unfortunately I don't access to any HP-UX machine where I can test it by myself. Have you compiled wget 1.12 on your machine? Thanks, Giuseppe
Re: [Bug-wget] Support of non-linux OS's going down the drain?
On 08/19/2011 12:18 AM, H.Merijn Brand wrote: With HP-UX 11.00 and HP C-ANSI-C it doesn't even *compile* anymore! (Re Support of non-linux OS's going down the drain?) If folks would like to see better support for non-GNU/Linux platforms, then folks using those platforms might do well to help test the software when test tarballs are released. There is one primary developer, and a smattering of sporadic contributors, and every time a release is made, the maintainer is always dismayed to find that all of a sudden people find all their particular platform bugs at that time. No one ever seems to pay much attention to the test packages that people make available, sometimes several months in advance, and the maintainer surely lacks the resources to test on many different platforms (particularly ones that don't lend themselves well to being run under a VM), so I'm afraid there's little that can be done about the problem. It wouldn't be appropriate to announce test source packages to as wide an audience as we do for full releases, and obviously the maintainer's own development platform (and whatever's most widely used among the mailing list users) will get the best treatment. (Not trying to criticize anyone or anything, just trying to make the point that the favored status of one particular OS is more or less inevitable.) -- Micah J. Cowan http://micah.cowan.name/
Re: [Bug-wget] Support of non-linux OS's going down the drain?
On Fri, 19 Aug 2011 12:06:03 -0700, Micah Cowan mi...@cowan.name wrote: bringing it back to the list ... You seem to think that I am angry or upset with you (or others) for some No, that just proves *I* sounded too harsch. reason. This is not so. And especially not for providing valuable feedback, for which I (and Giuseppe, I'm sure) am truly grateful. The only thing I was taking issue to was your subject heading, Support of non-linux OS's going down the drain?, because I think it's a bit unfair to judge support for OSses that are hard for us to test, and that no one wants to test during development. That might have been influenced by very recent experiences in many other projects that seems to prove that especially *young* (but enthusiastic) programmers forget that OS's other than Linux might not have available what their favourite Linux distro has available by default. A new dependency is very very costly. Neither am I upset that no one tests the alpha tarball packages, as seems to be your impression. I'm just pointing out that, since no one does, the testing then must happen after release, and so there's normally a series of patch releases following shortly afterwards, to improve support for less often-tested OSes. This has been true for every recent major release, as far as I can tell. It doesn't mean that support for other OSses has gone down the drain; just means it needs a little time to catch up. Ah right. There I indeed misinterpreted your original reply. Thanks for your help, and please do keep doing your part to help improve wget. -mjc On 08/19/2011 11:38 AM, H.Merijn Brand wrote: On Fri, 19 Aug 2011 09:48:12 -0700, Micah Cowanmi...@cowan.name wrote: On 08/19/2011 12:18 AM, H.Merijn Brand wrote: With HP-UX 11.00 and HP C-ANSI-C it doesn't even *compile* anymore! (Re Support of non-linux OS's going down the drain?) I fully understand below statement, but I do not feel guilty. I am close to full-time perl-tester. I invented the core-smoking process, that helps the perl5 development team to acquire test results for most architectures and operating systems without the testers actually being involved (though most testers actually get involved or get more interested than originally planned, which is a good thing). Our test bed runs on AIX, HP-UX, Linux, Windows and even OpenVMS and gives us a day-to-day overview of what patches break what. Note however that there are many many opensource projects, and that none of us have enough CPU cycles available to test more than one or two projects actively. I'm a perl5 committer, which makes me someone dedicated to perl5, but with a very clear and open view to problems other projects have on less active or available OS versions. That said, I will always feedback the problems and sometimes the solutions that I encounter in other projects. I'll give some example from the very recent past that took quite a bit of my precious free time. 1. Together with someone from the GNU gcc team we tracked down a bug in the assembler part of gcc itself that caused gcc to not only generate invalid assembly, but eventually caused gcc itself to crash in certain C code. Now fixed, will be part of gcc-4.6.2 2. SQLite code would not compile on HP-UX 11.00 and older under certain conditions. Together with some others, we analyzed the problem and found a fix 3. I gave a patch to the authors of ccache to enable running on both HP-UX and AIX. 4. I implemented (longer ago now) a way to allow grep to use PCRE in regular expressions. That is now - heavily changed - available as -P option 5. I helped implementing PCRE patterns in 'less' search algorithms 6. I gave a lot of feedback to the people that create and maintain git. It now runs out of the box on most HP-UX. These are all actions done together with others in tight cooperation. Feedback is vital. I might be the last one on this planet to still have *recent* opensource projects as prebuilt binary software depots available for most versions of HP-UX that are abandoned by HP. If I look at the amount of downloads for e.g. HP-UX 10.20, which I consider long dead, I see a proof that we -as open software community - make an awful lot of people very very happy. Help comes in grades. Feedback is one form of them. I certainly feel your angry when you say that others (including me) did not test pre-releases or test releases. I however just don't have the time to test each and every opensource project I ever use *for every release*. I mean, I just use wget. I used version 1.10/1 for a long time, until I ran into a problem. A new version is available, and I download it, run the documented build process (configuration, make, make test, make install) for those projects, and I report back what I find. If theere is an easy fix,