[fossil-users] forum-v2 can't compile

[Dingyuan Wang]

Hi,

I want to try out the forum-v2 branch, but it doesn't compile:

cc -I. -I./src -Ibld -Wall -DFOSSIL_ENABLE_JSON
-DFOSSIL_ENABLE_LEGACY_MV_RM=1 -DFOSSIL_ENABLE_TH1_DOCS
-DFOSSIL_ENABLE_TH1_HOOKS -DFOSSIL_DYNAMIC_BUILD=1  -DFOSSIL_HAVE_FUSEFS
 -g -O2 -DHAVE_AUTOCONFIG_H -D_HAVE_SQLITE_CONFIG_H -o bld/db.o -c bld/db_.c
./src/db.c: In function ‘db_err’:
./src/db.c:86:7: error: ‘rc’ undeclared (first use in this function)
   rc = 0 /* avoid HTTP 500 */;
   ^~
./src/db.c:86:7: note: each undeclared identifier is reported only once
for each function it appears in
___
fossil-users mailing list
fossil-users@lists.fossil-scm.org
http://lists.fossil-scm.org:8080/cgi-bin/mailman/listinfo/fossil-users

[fossil-users] Segfault on json test

[Dingyuan Wang]

Hi,

The following test:

  tclsh test/tester.tcl fossil json

fails with:

[...]
/home/gumble/dev/fossil/fossil json --json-input anon-1
ERROR: child killed: segmentation violation
JSON ERROR: unexpected token ":" at position 1; expecting VALUE
test json-login-a-env-validJSON FAILED!
test test-framework-json FAILED!
! json: key "payload" not known in dictionary
-code   = 1
-level  = 0
-errorstack = INNER dictGet CALL {test_json_payload json-login-a
{authToken name capabilities loginCookieName} {}}
-errorcode  = TCL LOOKUP DICT payload
-errorinfo  = key "payload" not known in dictionary
while executing
"dict get $::JR payload"
(procedure "test_json_payload" line 2)
invoked from within
"test_json_payload json-login-a {authToken name capabilities
loginCookieName} {}"
(file "/home/gumble/dev/fossil/test/json.test" line 263)
invoked from within
"source $testdir/$testfile.test"
-errorline  = 1"
* End of json: 3 errors so far **
* Final results: 3 errors out of 121 tests
* Considered failures: json-1 json-login-a-env-validJSON
test-framework-json
* Ignored results: 0 ignored errors out of 121 tests

The stack trace shows:

#0  0x55d0cdf3c39e in cgi_set_cookie (
zName=zName@entry=0x55d0ceaad160 "fossil-5617be7b37c6ab30",
zValue=zValue@entry=0x55d0ceaabaa0
"97fd92d7871e141066b93aedca178577682f732f/2458305.14815368/anonymous",
zPath=0x0, lifetime=lifetime@entry=21600)
at ./src/cgi.c:216
216 if( zPath[0]==0 ) zPath = "/";
#1  0x55d0cdf835e2 in login_set_anon_cookie (zIpAddr=,
zIpAddr@entry=0x0, zCookieDest=zCookieDest@entry=0x7ffef7822f88)
at ./src/login.c:329
#2  0x55d0cdf7daba in json_page_login () at ./src/json_login.c:151
#3  0x55d0cdf7a89f in json_dispatch_root_command (
zCommand=0x55d0ceac516c "login") at ./src/json.c:2209
#4  0x55d0cdf7abab in json_cmd_top () at ./src/json.c:2313
#5  0x55d0cdf2bb9e in main (argc=, argv=)
at ./src/main.c:760

Therefore, this commit
 https://www.fossil-scm.org/index.html/info/0a1f4ed6aa92f130
introduces the problem.
___
fossil-users mailing list
fossil-users@lists.fossil-scm.org
http://lists.fossil-scm.org:8080/cgi-bin/mailman/listinfo/fossil-users

Re: [fossil-users] email testing - no subscriber table?

[Dingyuan Wang]

Yes.

2018-06-26 02:28, Richard Hipp:
> On 6/24/18, Dingyuan Wang  wrote:
>>
>> I just got 50 exactly same "[fossil-src] activity alert" emails
>> describing 5 check-ins in three minutes. What's happening to the server?
>>
> 
> Did today's digest arrive successfully, and only once?
> 
___
fossil-users mailing list
fossil-users@lists.fossil-scm.org
http://lists.fossil-scm.org:8080/cgi-bin/mailman/listinfo/fossil-users

Re: [fossil-users] email testing - no subscriber table?

[Dingyuan Wang]

Hello,

I previously set to receive these before unsubscribing:

[x] Announcements
[x] Check-ins
[ ] Ticket changes
[x] Wiki
[x] Daily digest only

I just got 50 exactly same "[fossil-src] activity alert" emails
describing 5 check-ins in three minutes. What's happening to the server?

Besides the mail server problem, this subject is too general. Something
like "Daily digest: 5 check-ins by drh[, usera, userb]" is better.

2018-06-24 18:27, j. van den hoff:
> 
> 
> On Sun, 24 Jun 2018 12:22:07 +0200, Richard Hipp  wrote:
> 
>> On 6/24/18, j. van den hoff  wrote:
>>> On Sun, 24 Jun 2018 12:08:30 +0200, Richard Hipp  wrote:
>>>
 The UPDATE syntax error should be fixed now.  Please try it again.
>>>
>>> yes, it works now. thank you. NB: I received the notification email
>>> regarding the fix prior to actually confirming the subscription again --
>>> so my confirmation went through despite the SQL error? or is this
>>> indication of some flaw in the subscription logic?
>>>
>>
>> You had already been verify by the fact of visiting the page itself.
>> That was a separate UPDATE (with the correct syntax!) that was
>> performed in a different transaction.
>>
>> Thanks for pointing this out, though.  I did have to think about it
>> for a few seconds.
> 
> thanks for this explanation...
> 
> regarding the format/wording of the notification, it currently reads:
> 
> 8<--
> This is an automated email sent by the Fossil repository at
> https://fossil-scm.org/fossil to report changes.
> 
> == 2018-06-24 10:07:29 Check-In ==
> Fix an SQL syntax error. (user: drh tags: trunk)
> https://fossil-scm.org/fossil/info/0398e41aa6f72c0da60b
> 
> 
> Subscription info: https://fossil-scm.org/fossil/alerts/*
> 8<--
> 
> personally, I would prefer to simply omit the first line or to put it at
> the bottom of the footer, since I would recognize the nature of the mail
> without this explanation, so it's essentially explaining the obvious and
> a minor distraction from the actual content of the mail (and the repos
> identity can also be taken from the `subscription info' link).
> 
> additionally, mabye shorten the footer separation line to exactly two
> `--', treating the footer as the sender's "affiliation/identity" (which
> usually leads to a less prominent display by the email client).
>>
> 
> 
___
fossil-users mailing list
fossil-users@lists.fossil-scm.org
http://lists.fossil-scm.org:8080/cgi-bin/mailman/listinfo/fossil-users

Re: [fossil-users] How to rename a directory

[Dingyuan Wang]

The trunk version. The previous implementation is actually broken except
for file renaming.

Index: src/add.c
==
--- src/add.c
+++ src/add.c
@@ -859,10 +859,12 @@
   int vid;
   int moveFiles;
   int dryRunFlag;
   int softFlag;
   int hardFlag;
+  int origType;
+  int destType;
   char *zDest;
   Blob dest;
   Stmt q;

   db_must_be_within_tree();
@@ -900,15 +902,23 @@
 "UPDATE vfile SET origname=pathname WHERE origname IS NULL;"
   );
   db_multi_exec(
 "CREATE TEMP TABLE mv(f TEXT UNIQUE ON CONFLICT IGNORE, t TEXT);"
   );
-  if( file_isdir(zDest, RepoFILE)!=1 ){
+  if( g.argc!=4 ){
+origType = -1;
+  }else{
+origType = (file_isdir(g.argv[2], RepoFILE) == 1);
+  }
+  destType = file_isdir(zDest, RepoFILE);
+  if( origType==-1 && destType!=1 ){
+usage("OLDNAME NEWNAME");
+  }else if( origType==1 && destType==2 ){
+fossil_fatal("cannot rename '%s' to '%s' since another file named"
+ " '%s' exists", g.argv[2], zDest, zDest);
+  }else if( origType==0 && destType!=1 ){
 Blob orig;
-if( g.argc!=4 ){
-  usage("OLDNAME NEWNAME");
-}
 file_tree_name(g.argv[2], , 0, 1);
 db_multi_exec(
   "INSERT INTO mv VALUES(%B,%B)", , 
 );
   }else{
@@ -936,10 +946,12 @@
 const char *zPath = db_column_text(, 0);
 int nPath = db_column_bytes(, 0);
 const char *zTail;
 if( nPath==nOrig ){
   zTail = file_tail(zPath);
+}else if( destType==1 ){
+  zTail = [nOrig-strlen(file_tail(zOrig))];
 }else{
   zTail = [nOrig+1];
 }
 db_multi_exec(
   "INSERT INTO mv VALUES('%q','%q%q')",

Index: test/mv-rm.test
==
--- test/mv-rm.test
+++ test/mv-rm.test
@@ -16,14 +16,10 @@
 
 #
 # MV / RM Commands
 #

-set path [file dirname [info script]]
-
-require_no_open_checkout
-
 
 # Setup: Add Files and Commit  #
 

 test_setup; set rootDir [file normalize [pwd]]
@@ -413,21 +409,22 @@
 # Test 18: Move Directory to New Directory #
 

 fossil mv --hard subdirC subdirD
 test mv-file-new-directory-7 {
-  [normalize_result] eq "RENAME subdirC subdirD\nMOVED_FILE
${rootDir}/subdirC"
+  [normalize_result] eq "RENAME subdirC/f10 subdirD/f10\nRENAME
subdirC/f11 subdirD/f11\nMOVED_FILE ${rootDir}/subdirC/f10\nMOVED_FILE
${rootDir}/subdirC/f11"
 }

 test mv-file-new-directory-8 {[file size subdirD/f10] == 3}
 test mv-file-new-directory-9 {[read_file subdirD/f10] eq "f10"}
 test mv-file-new-directory-10 {[file size subdirD/f11] == 3}
 test mv-file-new-directory-11 {[read_file subdirD/f11] eq "f11"}

 fossil revert
+puts [normalize_result]
 test mv-file-new-directory-12 {
-  [normalize_result] eq "REVERT   subdirC/f10\nREVERT
subdirC/f11${undoMsg}"
+  [normalize_result] eq "DELETE   subdirD/f10\nDELETE
subdirD/f11\nREVERT   subdirC/f10\nREVERT   subdirC/f11${undoMsg}"
 }

 test mv-file-new-directory-13 {[file size subdirC/f10] == 3}
 test mv-file-new-directory-14 {[read_file subdirC/f10] eq "f10"}
 test mv-file-new-directory-15 {[file size subdirC/f11] == 3}



2018-05-06 01:25, Richard Hipp:> On 5/5/18, Dingyuan Wang
<gumb...@aosc.io> wrote:
>>
>> The fossil mv command seems can't rename a directory.
> 
> I thought somebody had fixed that.  But maybe not.  What version are you 
> using?
> 
> Renaming directories, or even files, is not something that I do very
> often, as I find that it disrupts the history of the project, making
> it more difficult to answer questions like "why and when was this line
> of code inserted?"  If the code keeps moving around, it becomes more
> difficult to track which file that line of code started in.  To
> illustrate the infrequency with which I rename files, here is a
> complete list of file rename operations that have occurred on the
> SQLite project during the 8.7 years that it has been tracked by
> Fossil:
> 
> https://sqlite.org/src/test-rename-list
> 
> That's a short list when you consider that there have been 13616
> check-ins in that same time period.
> 
> The above sounds like I'm making excuses.  I shouldn't be.  Fossil
> *should* rename the directory for you without giving you any grief.
> It should "just work".  I'm sorry that didn't.  That needs to be
> fixed.  (You can submit patches if you like :-))  The purpose of my
> long reply is to point out that renaming is not a frequently used
> feature, and is thus more likely to contain shortcoming and bugs,
> which you have apparently now discovered.
> 
___
fossil-users mailing list
fossil-users@lists.fossil-scm.org
http://lists.fossil-scm.org:8080/cgi-bin/mailman/listinfo/fossil-users

[fossil-users] How to rename a directory

[Dingyuan Wang]

Hi,

The fossil mv command seems can't rename a directory.
eg.
`fossil mv --hard test test1` then `fossil status` will give
`RENAMEDtest1`.

Suppose that there is a file `a/test`. `fossil mv --hard a b` then
`fossil status` will give `MISSINGa/test`. `fossil commit` will
abort due to missing files.

It seems that fossil didn't record that the files in the directory has
been moved.
___
fossil-users mailing list
fossil-users@lists.fossil-scm.org
http://lists.fossil-scm.org:8080/cgi-bin/mailman/listinfo/fossil-users

[fossil-users] import multiple branch names from git

[Dingyuan Wang]

Hi,

When exporting git repos with multiple branches pointing to one commit,
fossil only tag one of the branches, because fossil doesn't recognize
"reset" commands.

For example, this fast-export:

commit refs/heads/bugfix
mark :83209
author Someone  1523846129 +
committer Someone  1523938423 +
[...]

reset refs/heads/staging
from :83209

reset refs/pull/74/head
from :8557

reset refs/heads/staging
from :83209

reset refs/heads/master
from :83209

The last commit :83209 has three branch heads (bugfix, staging, master).
Fossil only recognized the branch written in the "commit" command.
___
fossil-users mailing list
fossil-users@lists.fossil-scm.org
http://lists.fossil-scm.org:8080/cgi-bin/mailman/listinfo/fossil-users

[fossil-users] patch: support using author in git fast-export

[Dingyuan Wang]

Hi,

The git fast-import format can specify different 'author' and
'committer', while fossil only uses one. This patch adds the import
option --use-author to set 'author' as the checkin user, instead of
'committer'. This avoid problems when 'committer' is not meaningful, eg.
'GitHub '. When --use-author is on and there is no
'author', 'committer' is used as fallback.


Index: src/import.c
==
--- src/import.c
+++ src/import.c
@@ -517,10 +517,11 @@
 }


 static struct{
   const char *zMasterName;/* Name of master branch */
+  int authorFlag; /* Use author as checkin committer */
 } ggit;

 /*
 ** Read the git-fast-import format from pIn and insert the corresponding
 ** content into the database.
@@ -619,19 +620,23 @@
   gg.aData = 0;
   gg.nData = 0;
 }
   }
 }else
-if( strncmp(zLine, "author ", 7)==0 ){
+if( (!ggit.authorFlag && strncmp(zLine, "author ", 7)==0)
+|| (ggit.authorFlag && strncmp(zLine, "committer ",10)==0
+&& gg.zUser!=NULL) ){
   /* No-op */
 }else
 if( strncmp(zLine, "mark ", 5)==0 ){
   trim_newline([5]);
   fossil_free(gg.zMark);
   gg.zMark = fossil_strdup([5]);
 }else
-if( strncmp(zLine, "tagger ", 7)==0 || strncmp(zLine, "committer
",10)==0 ){
+if( strncmp(zLine, "tagger ", 7)==0
+|| (ggit.authorFlag && strncmp(zLine, "author ", 7)==0)
+|| strncmp(zLine, "committer ",10)==0 ){
   sqlite3_int64 secSince1970;
   z = strchr(zLine, ' ');
   while( fossil_isspace(*z) ) z++;
   if( (zTo=strchr(z, '>'))==NULL ) goto malformed_line;
   *(++zTo) = '\0';
@@ -1603,10 +1608,11 @@
 **   --gitImport from the git-fast-export file format (default)
 **Options:
 **  --import-marks  FILE Restore marks table from FILE
 **  --export-marks  FILE Save marks table to FILE
 **  --rename-master NAME Renames the master branch to NAME
+**  --use-authorUses author as the committer
 **
 **   --svnImport from the svnadmin-dump file format.  The default
 **behaviour (unless overridden by --flat) is to treat 3
 **folders in the SVN root as special, following the
 **common layout of SVN repositories.  These are (by
@@ -1727,10 +1733,11 @@
 markfile_in = find_option("import-marks", 0, 1);
 markfile_out = find_option("export-marks", 0, 1);
 if( !(ggit.zMasterName = find_option("rename-master", 0, 1)) ){
   ggit.zMasterName = "master";
 }
+ggit.authorFlag = find_option("use-author", 0, 0)!=0;
   }
   verify_all_options();

   if( g.argc!=3 && g.argc!=4 ){
 usage("--git|--svn ?OPTIONS? NEW-REPOSITORY ?INPUT-FILE?");

___
fossil-users mailing list
fossil-users@lists.fossil-scm.org
http://lists.fossil-scm.org:8080/cgi-bin/mailman/listinfo/fossil-users

[fossil-users] patch: fossil import sha3 marks support

[Dingyuan Wang]

Hi,

The patch fixes the problem that fossil doesn't recognize sha3 marks
file. eg.

first do

$ mkdir .marks
$ git fast-export --all --export-marks=.marks/git | fossil import
--export-marks .marks/fossil --git repo.fossil

then

$ git fast-export --all --import-marks=.marks/git
--export-marks=.marks/git | fossil import --import-marks .marks/fossil
--export-marks .marks/fossil --git --incremental repo.fossil


Index: src/export.c
==
--- src/export.c
+++ src/export.c
@@ -35,16 +35,16 @@
 **   and fossil commits.
 **   -git_name: This is the mark name that identifies the commit to git.
 **  It will always begin with a ':'.
 **   -rid: The unique object ID that identifies this commit within the
 ** repository database.
-**   -uuid: The SHA-1 of artifact corresponding to rid.
+**   -uuid: The SHA-1/SHA-3 of artifact corresponding to rid.
 */
 struct mark_t{
   char *name;
   int rid;
-  char uuid[41];
+  char uuid[65];
 };
 #endif

 /*
 ** Output a "committer" record for the given user.
@@ -340,22 +340,23 @@
   }else{
 mark->name = fossil_strdup(cur_tok);
   }

   cur_tok = strtok(NULL, "\n");
-  if( !cur_tok || strlen(cur_tok)!=40 ){
+  if( !cur_tok || (strlen(cur_tok)!=40 && strlen(cur_tok)!=64) ){
 free(mark->name);
-fossil_trace("Invalid SHA-1 in marks file: %s\n", cur_tok);
+fossil_trace(
+  "Invalid SHA-1/SHA-3 in marks file: %d %s\n", strlen(cur_tok),
cur_tok);
 return -1;
   }else{
 sqlite3_snprintf(sizeof(mark->uuid), mark->uuid, "%s", cur_tok);
   }

   /* make sure that rid corresponds to UUID */
   if( fast_uuid_to_rid(mark->uuid)!=mark->rid ){
 free(mark->name);
-fossil_trace("Non-existent SHA-1 in marks file: %s\n", mark->uuid);
+fossil_trace("Non-existent SHA-1/SHA-3 in marks file: %s\n",
mark->uuid);
 return -1;
   }

   /* insert a cross-ref into the 'xmark' table */
   insert_commit_xref(mark->rid, mark->name, mark->uuid);

___
fossil-users mailing list
fossil-users@lists.fossil-scm.org
http://lists.fossil-scm.org:8080/cgi-bin/mailman/listinfo/fossil-users

[fossil-users] 413 Request Entity Too Large

[Dingyuan Wang]

Hi,

When I committed a large blob, I can't sync with my server, which uses
nginx for reverse proxy.
"server says: 413 Request Entity Too Large"

I've set the max-upload to 1000 but it doesn't help.
___
fossil-users mailing list
fossil-users@lists.fossil-scm.org
http://lists.fossil-scm.org:8080/cgi-bin/mailman/listinfo/fossil-users

[fossil-users] How to find deleted files in web interface

[Dingyuan Wang]

Hi,

In the command line, fossil timeline -p somefile can show the commit
when the file is deleted, while web interface /finfo?name=somefile
can't. It also don't show merges.

What's the web interface equivalence of `timeline` for a file, or how
can I find when the file is deleted?
___
fossil-users mailing list
fossil-users@lists.fossil-scm.org
http://lists.fossil-scm.org:8080/cgi-bin/mailman/listinfo/fossil-users

[fossil-users] Unmatched parenthesis in the new file history

[Dingyuan Wang]

Hi,

The recent CSS update looks good. There is a bug in "File History" view.

eg.


The commit metadata says:

file: [92589049] check-in: [e59a7fd3] user: drh, branch: trunk, size: 28652)

This appears in each kind of view.

___
fossil-users mailing list
fossil-users@lists.fossil-scm.org
http://lists.fossil-scm.org:8080/cgi-bin/mailman/listinfo/fossil-users

Re: [fossil-users] Initial empty checkin?

[Dingyuan Wang]

Fossil import now has a bug that it makes a repo with SHA3 hashes, but
can't do incremental updates because of the SHA3 hashes. My workaround
is to first `fossil new --sha1` and then import. The initial empty
checkin is annoying for importing.

2017-08-17 18:32, Stephan Beal:
> Fwiw: My recollection and opinion are the same.
> 
> - stephan
> Sent from a mobile device, possibly from bed. Please excuse brevity,
> typos, and top-posting.
> 
> On Aug 17, 2017 07:35, "Andy Bradford"  > wrote:
> 
> Thus said Andy Goth on Wed, 16 Aug 2017 10:47:56 -0500:
> 
> > What of this old thread? Are  the issues it discusses still pertinent,
> > or have they been resolved?
> 
> I believe  all the  relevant issues were  actually resolved,  however, I
> think it  was still unfavorable  given the  time that was  available for
> testing the changes before release.
> 
> Personally,  I would  be  in  favor of  retaining  the  old behavior  by
> default,  but allowing  a command  line option  to ``fossil  init'' that
> would initialize a new repository without it.
> 
> Andy
> --
> TAI64 timestamp: 400059952b4d
> 
> 
> ___
> fossil-users mailing list
> fossil-users@lists.fossil-scm.org
> 
> http://lists.fossil-scm.org:8080/cgi-bin/mailman/listinfo/fossil-users
> 
> 
> 
> 
> ___
> fossil-users mailing list
> fossil-users@lists.fossil-scm.org
> http://lists.fossil-scm.org:8080/cgi-bin/mailman/listinfo/fossil-users
> 
___
fossil-users mailing list
fossil-users@lists.fossil-scm.org
http://lists.fossil-scm.org:8080/cgi-bin/mailman/listinfo/fossil-users

[fossil-users] Bug: can't import to SHA3 repos

[Dingyuan Wang]

Dear all,

In the latest version, Fossil uses a SHA3 hash by default. When I'm
trying to sync from Git to Fossil, it refuses because of the SHA3 in the
marks file.

First:

$ git fast-export --all --import-marks=../somerepo.git-marks
--export-marks=../somerepo.git-marks | fossil import --import-marks
../somerepo.fossil-marks --export-marks ../somerepo.fossil-marks --git
../somerepo.fossil

After some commits in the git repo:

$ git fast-export --all --import-marks=../somerepo.git-marks
--export-marks=../somerepo.git-marks | ~/dev/fossil/fossil import
--import-marks ../somerepo.fossil-marks --export-marks
../somerepo.fossil-marks --git --incremental ../somerepo.fossil
Invalid SHA-1 in marks file:
fb5191794c10f2b66ff915fff4e40964fad10754b9ad47c6a7a5cad395aec93d
error importing marks from file: ../somerepo.fossil-marks
___
fossil-users mailing list
fossil-users@lists.fossil-scm.org
http://lists.fossil-scm.org:8080/cgi-bin/mailman/listinfo/fossil-users