Ticket #3243: 1242387095.dpatch

File 1242387095.dpatch, 5.7 KB (added by nwn, 8 years ago)

patch to this bug, It works fine and passes tests

1Fri May 15 20:29:51 JST 2009  Yusaku Hashimoto <nonowarn@gmail.com>
2  * before fork(), stopTimer() for child process, after fork(), startTimer() for parent process
4New patches:
6[before fork(), stopTimer() for child process, after fork(), startTimer() for parent process
7Yusaku Hashimoto <nonowarn@gmail.com>**20090515112951] {
8hunk ./cbits/runProcess.c 66
9     // shared between parent and child), and the parent behaves as if
10     // the signal had been raised.
11     blockUserSignals();
12+       stopTimer();
14     switch(pid = fork())
15     {
16hunk ./cbits/runProcess.c 71
17     case -1:
18+               startTimer();
19         unblockUserSignals();
20         if (fdStdIn == -1) {
21             close(fdStdInput[0]);
22hunk ./cbits/runProcess.c 189
23         }
24        break;
25     }
26+       startTimer();
27     unblockUserSignals();
29     return pid;
34[Remove some incorrect (out of date?) documentation; fixes trac #3152
35Ian Lynagh <igloo@earth.li>**20090410173535]
36[fix this test for MSYS
37Simon Marlow <marlowsd@gmail.com>**20090310130059]
38[Avoid vfork() bear traps
39Simon Marlow <marlowsd@gmail.com>**20090309161943
40 Ignore-this: db3593f7e5db1db560842c729e47c551
41 We can't call setIOManagerPipe() in the vfork() child, because the
42 change will be reflected in the parent too.  Add a large warning to
43 that effect.
45 I tried changing vfork() to fork(), but it seems that this leads to a
46 different problem: the fork() sometimes takes so long that it gets
47 repeatedly interrupted by the timer signal and never makes progress.
48 I could disable the timer for a while, but decided to back off and fix
49 the vfork() version instead.
51[fix documentation for termianteProcess (#2954)
52Simon Marlow <marlowsd@gmail.com>**20090305145745
53 Ignore-this: 341860d041d9dd7345f41dec64de0be7
55[Add config.guess, config.sub and install-sh
56Ian Lynagh <igloo@earth.li>**20090307153856]
57[Fix #2870: User signals are not blocked before 'fork' in runInteractiveProcess
58Simon Marlow <marlowsd@gmail.com>**20090219123235
59 Ignore-this: dd132c40e9ccefc8362caf44216a81a0
61[Require Cabal version >= 1.6
62Ian Lynagh <igloo@earth.li>**20090122011323]
63[Add "bug-reports" and "source-repository" info to the Cabal file
64Ian Lynagh <igloo@earth.li>**20090121182651]
65[Avoid using IOError internals
66Ian Lynagh <igloo@earth.li>**20090104173211]
67[fix warning on Windows
68Simon Marlow <marlowsd@gmail.com>**20081114124114]
69[add some extra_clean
70Simon Marlow <marlowsd@gmail.com>**20081114104354]
72Simon Marlow <marlowsd@gmail.com>**20081114104329]
73[Pass the correct name to runGenProcess_ for sake of error messages
74Duncan Coutts <duncan@haskell.org>**20081113105602
75 Otherwise error messages from calling createProcess will report the
76 internal name runGenProcess_ which is misleading. Similarly for the
77 system and rawSystem functions.
79[remove non-existent autoconf files from extra-source-files
80Simon Marlow <marlowsd@gmail.com>**20081112125459]
81[fix .cabal style bugs noticed by Cabal 1.6
82Simon Marlow <marlowsd@gmail.com>**20081105100456]
83[a Windows fix, and a base-4 fix
84Simon Marlow <marlowsd@gmail.com>**20081024135423]
85[improve docs for terminateProcess (#2638)
86Simon Marlow <marlowsd@gmail.com>**20081024134725
87 When used with a shell command on Windows, terminateProcess only kills
88 the shell, not the command.
90[Make this compile with ghc-6.8.3; update version to
91Simon Marlow <marlowsd@gmail.com>**20081023143737]
92[Use printf rather than echo for the process008 test
93Ian Lynagh <igloo@earth.li>**20081001232339
94 It doesn't work on Windows for me otherwise
96[Bump version number to
97Ian Lynagh <igloo@earth.li>**20080920160210]
98[TAG 6.10 branch has been forked
99Ian Lynagh <igloo@earth.li>**20080919123438]
100[follow library changes
101Ian Lynagh <igloo@earth.li>**20080903223613]
102[non-GHC: export System.Process, not System.Process.Internals
103Ross Paterson <ross@soi.city.ac.uk>**20080831181126]
104[fix cabal build-depends for nhc98
106[add category field
107Ross Paterson <ross@soi.city.ac.uk>**20080824003013]
108[add extra-source-files
109Ross Paterson <ross@soi.city.ac.uk>**20080824002709]
110[We now depend on concurrent (split off from base)
111Ian Lynagh <igloo@earth.li>**20080824135154]
112[use fmap instead of importing Applicative
113Ross Paterson <ross@soi.city.ac.uk>**20080817001800
115 Importing Applicative into this module causes the instances in
116 Control.Monad.Instances to leak into the Haskell 98 module System,
117 which would break Haskell 98 programs that define these instances.
119[switch to new exceptions
120Ross Paterson <ross@soi.city.ac.uk>**20080814150115]
121[Windows fix
122Ian Lynagh <igloo@earth.li>**20080803180419]
123[We don't need to explicitly "import Prelude"
124Ian Lynagh <igloo@earth.li>**20080803152344
125 We had to do this beofre to ensure that things built in the right order,
126 but now Prelude is in a different package.
128[Follow extensible exceptions changes
129Ian Lynagh <igloo@earth.li>**20080623193127]
130[Silence warnings
131Ian Lynagh <igloo@earth.li>**20080703154751]
132[Fix warnings
133Ian Lynagh <igloo@earth.li>**20080620011826]
134[Avoid using deprecated flags
135Ian Lynagh <igloo@earth.li>**20080616145419]
136[turn off new functions for Hugs (for now)
137Ross Paterson <ross@soi.city.ac.uk>**20080615224631]
138[try to make this test a bit more portable (don't depend on the output of ls)
139Simon Marlow <marlowsd@gmail.com>**20080604092520]
140[update Windows output
141Simon Marlow <marlowsd@gmail.com>**20080529153428]
142[fix bug on Windows when redirecting stderr
143Simon Marlow <marlowsd@gmail.com>**20080529153316]
144[readProcessWithExitCode now returns separate stdout & stderr
145Simon Marlow <marlowsd@gmail.com>**20080528103212]
146[TAG 2008-05-28
147Ian Lynagh <igloo@earth.li>**20080528004421]
148Patch bundle hash: