Many hyperlinks are disabled.
Use anonymous login
to enable hyperlinks.
46 check-ins related to "win32-longpath"
2014-06-12
| ||
17:26 | merge trunk Leaf check-in: e751da62cb user: jan.nijtmans tags: win32-longpath | |
17:25 | Version 1.29 check-in: 3e5ebe2b90 user: drh tags: trunk, release, version-1.29 | |
2014-04-14
| ||
13:03 | merge trunk. Add support for extended UNC paths as well. check-in: 9d4a73cb40 user: jan.nijtmans tags: win32-longpath | |
2014-04-11
| ||
21:30 | Added new /reports page: byweekday. Most tags and commits happen on Thursdays. Most tickets on Wednesdays. check-in: 2cf6304bc3 user: stephan tags: trunk | |
2014-03-13
| ||
21:06 | merge trunk check-in: d88d1bc0f6 user: jan.nijtmans tags: win32-longpath | |
13:17 | Backout [e6d3e878a9]: Clicking on timeline nodes for displaying a diff doesn't work any more with this change (Sorry, Stefan!). check-in: afcdc6ce81 user: jan.nijtmans tags: trunk | |
2014-02-14
| ||
10:25 | Merge in updates from trunk. Some implementation simplifications and more comments (mainly in the SQLite part). check-in: 9cc1c9d690 user: jan.nijtmans tags: win32-longpath | |
2014-02-13
| ||
15:17 | Take over "Fixes to the editline support" and "Updates to the command-line shell" from SQLite trunk, keeping the two in sync better. Except for the addition of the ".save" command in "fossil sqlite3", it has no effect. check-in: e327614047 user: jan.nijtmans tags: trunk | |
2014-01-23
| ||
21:47 | merge trunk check-in: abb2400434 user: jan.nijtmans tags: win32-longpath | |
21:18 | merge branch-1.28, a few other cherry-picks from trunk check-in: 303e7af7c3 user: jan.nijtmans tags: win32-longpath | |
10:29 | No longer require that the HOME directory is writable if the fossil configuration file is already there. (With some simplifications) check-in: b4d538f8c6 user: jan.nijtmans tags: trunk | |
2014-01-18
| ||
13:49 | Skip merge-utf-*.23 and merge-utf-*.32 test-cases, which currently fail. They are not crucial for the 1.28 release (in Fossil 1.26 and 1.27 they failed as well). Probably a problem in the test-case itself (some special character in test/utf.test?) check-in: 4b3b89b51e user: jan.nijtmans tags: branch-1.28 | |
2013-12-30
| ||
20:57 | merge trunk check-in: 02a0e8890e user: jan.nijtmans tags: win32-longpath | |
2013-12-29
| ||
00:53 | Fix floating-point divide by zero issue in TH1 reported on the mailing list and add tests. check-in: 479b3de1d2 user: mistachkin tags: trunk | |
2013-12-20
| ||
12:12 | The same, but then with modified SQLite (minimal patch to function winFullPathname this time). This way, applications wanting to use the "win32-longpath" VFS don't have to do tricky extended-path modifications any more. check-in: 8e8e3a11bc user: jan.nijtmans tags: win32-longpath | |
12:04 | Merge trunk. Some more comments. check-in: 4e3f4d26d8 user: jan.nijtmans tags: win32-longpath | |
09:36 | Update to SQLite snapshot check-in: 44dec8ff07 user: jan.nijtmans tags: trunk | |
2013-12-18
| ||
08:11 | Merge trunk. Valid extended paths must have >6 chars, e.g. "\\?\C:\" (+1 for closing '\0'), this way make sure that wUnicode always points to allocated memory. check-in: 1db9134d6b user: jan.nijtmans tags: win32-longpath | |
06:37 | Consistency fix for a Fossil link in the enhanced default skin. check-in: 04186fcc12 user: mistachkin tags: trunk | |
00:18 | Remove superfluous comment line in makemake.tcl. check-in: e977f86dfd user: mistachkin tags: win32-longpath | |
00:17 | Merge updates from trunk. check-in: ef9f707be1 user: mistachkin tags: win32-longpath | |
2013-12-17
| ||
22:31 | Further cleanup in the MSVC build. Resync the makefiles with the list of SQLite options. check-in: 61e70f1fc7 user: mistachkin tags: trunk | |
14:49 | Eliminate use of "goto". Fix UNC -> Extended UNC translation when path>260 chars. check-in: 83087c8df3 user: jan.nijtmans tags: win32-longpath | |
12:03 | It appears that in [4f0413381d], Makefile.msc was edited without corresponding change in makemake.tcl. Correct that, so this change will not be accidently reverted again (as I did in the previous commit) check-in: c6f255607b user: jan.nijtmans tags: win32-longpath | |
11:56 | Align recommended SQLite compilation options with actual options used when compiling SQLite. (I don't think SQLITE_ENABLE_EXPLAIN_COMMENTS should be one of them, it never was in earlier releases) check-in: d0814e59e2 user: jan.nijtmans tags: win32-longpath | |
10:56 | merge trunk check-in: 6a8ddc43fc user: jan.nijtmans tags: hidden-tag | |
10:51 | If 'newbranch' field only contains spaces, consider it empty. check-in: 5ceef9bcdd user: jan.nijtmans tags: win32-longpath | |
10:20 | Break the 260 character path barrier on Win32. Add test-case "win32-longpath" for proving that it works. check-in: 4b49bb9cc3 user: jan.nijtmans tags: win32-longpath | |
10:14 | Move special path handling from SQLite to fossil. check-in: 1dd5a50941 user: jan.nijtmans tags: win32-longpath | |
09:32 | If "fossil new" or "fossil open"'s argument has an extended path prefix like "\\?\C:\", don't process the '?' and ':' in it as special path characters, only do '/' -> '\' translation. This makes extended paths usable in fossil everywhere (hopefully), for now <260. check-in: 04f4e699d4 user: jan.nijtmans tags: win32-longpath | |
06:04 | Change the minimum length for wiki page names from 3 to 1. check-in: b7ff13a0aa user: joel tags: trunk | |
2013-12-16
| ||
15:12 | Hm. "catch" doesn't appear to work for reporting the win32-longpath error correctly, instead check whether CANTOPEN is part of the output. check-in: 4b4606d8b8 user: jan.nijtmans tags: win32-longpath | |
14:25 | Better clean-up after win32-longpath test. Revert accidental change in Makefile.mingw. check-in: 698a2d5968 user: jan.nijtmans tags: win32-longpath | |
13:37 | Fix bug in utf8.c (don't translate first ':' in extended path). Add test-case (win32-longpath.test). Some formatting (don't use tabs). check-in: 7d5c1bbd9c user: jan.nijtmans tags: win32-longpath | |
10:57 | Merge trunk. Eliminate need for special Makefile.Cygwin. Simplify handling of '\\?\' prefix: Only add it when necessary (calculated path>=260 chars). check-in: 23b849cce5 user: jan.nijtmans tags: win32-longpath | |
10:46 | If the name of a 'sym-<branchname>' tag sorts before the 'branch' tag, no option in the ci_edit page should be displayed to cancel the tag. Bug-fix discovered and implemented by Andy Bradford. check-in: 3529f2c47b user: jan.nijtmans tags: trunk | |
2013-12-13
| ||
13:53 | merge trunk check-in: 46e1d4c812 user: jan.nijtmans tags: win32-longpath | |
13:48 | Better use GetCurrentDirectoryW in stead of _wgetcwd. check-in: 4e463bf7ba user: jan.nijtmans tags: trunk | |
2013-12-12
| ||
16:13 | One more: GetFullPathNameW() needs the "\\?\"-prefix as well if the input path is already >MAX_PATH. But we don't want the resulting directory to have this prefix, so strip it off when present. check-in: e5fef7982d user: jan.nijtmans tags: win32-longpath | |
15:01 | Final bug-fix: Before prefixing a path with "\\?\", check if it is prefixed already. If so, don't do it again. Everything seems to work fine now! check-in: c952c773ac user: jan.nijtmans tags: win32-longpath | |
14:27 | Implement file_access() function such that it accepts paths>MAX_PATH. Implementation copied (with some simplifications) from Tcl 8.6 check-in: ba4b3ac1d2 user: jan.nijtmans tags: win32-longpath | |
13:54 | Fix fossil_stat() and fossil_chdir() such that they accept paths>MAX_PATH. For file_access(), implement a workaround that the '\\?\' prefix can be handled. check-in: 29f023fe53 user: jan.nijtmans tags: win32-longpath | |
11:16 | Bug-fix: didn't compile on win32, and handle extended UNC paths correctly. check-in: da8d516fe1 user: jan.nijtmans tags: win32-longpath | |
09:37 | Make "win32-longpath" the default VFS on win32, eliminating all path limitations (up to ~32767 chars). TODO: eliminate use of the the function _wstati64(), that appears to be the only Win32 function left which cannot handle such long paths. Everything else needed is done. check-in: 1b9893bdc8 user: jan.nijtmans tags: win32-longpath | |
08:43 | Put back 'win32-longpath' VFS as default for Cygwin: Cygwin doesn't suffer from path limitations (at least, not until 4096 chars), so usage of the 'win32' VFS is not reasonable as default. check-in: a76039815d user: jan.nijtmans tags: trunk | |
2013-12-11
| ||
13:00 | As it turns out that _wstati64() cannot handled the special "\\?\" prefix, work around that. Otherwise the win32-longpath VFS is quite useless for fossil. Maybe a better solution should be worked out, not using _wstati64() at all. check-in: bb440899d3 user: jan.nijtmans tags: trunk | |