Mercurial > hg > xemacs-beta
comparison src/callproc.c @ 363:972bbb6d6ca2 r21-1-11
Import from CVS: tag r21-1-11
| author | cvs |
|---|---|
| date | Mon, 13 Aug 2007 10:59:28 +0200 |
| parents | 8e84bee8ddd0 |
| children | 30d2cfa1092a |
comparison
equal
deleted
inserted
replaced
| 362:1e474c183006 | 363:972bbb6d6ca2 |
|---|---|
| 609 #ifdef SET_EMACS_PRIORITY | 609 #ifdef SET_EMACS_PRIORITY |
| 610 if (emacs_priority != 0) | 610 if (emacs_priority != 0) |
| 611 nice (- emacs_priority); | 611 nice (- emacs_priority); |
| 612 #endif | 612 #endif |
| 613 | 613 |
| 614 /* Under Windows, we are not in a child process at all, so we should | |
| 615 not close handles inherited from the parent -- we are the parent | |
| 616 and doing so will screw up all manner of things! Similarly, most | |
| 617 of the rest of the cleanup done in this function is not done | |
| 618 under Windows. | |
| 619 | |
| 620 #### This entire child_setup() function is an utter and complete | |
| 621 piece of shit. I would rewrite it, at the very least splitting | |
| 622 out the Windows and non-Windows stuff into two completely | |
| 623 different functions; but instead I'm trying to make it go away | |
| 624 entirely, using the Lisp definition in process.el. What's left | |
| 625 is to fix up the routines in event-msw.c (and in event-Xt.c and | |
| 626 event-tty.c) to allow for stream devices to be handled correctly. | |
| 627 There isn't much to do, in fact, and I'll fix it shortly. That | |
| 628 way, the Lisp definition can be used non-interactively too. */ | |
| 614 #if !defined (NO_SUBPROCESSES) && !defined (WINDOWSNT) | 629 #if !defined (NO_SUBPROCESSES) && !defined (WINDOWSNT) |
| 615 /* Close Emacs's descriptors that this process should not have. */ | 630 /* Close Emacs's descriptors that this process should not have. */ |
| 616 close_process_descs (); | 631 close_process_descs (); |
| 617 #endif /* not NO_SUBPROCESSES */ | 632 #endif /* not NO_SUBPROCESSES */ |
| 633 #ifndef WINDOWSNT | |
| 618 close_load_descs (); | 634 close_load_descs (); |
| 635 #endif | |
| 619 | 636 |
| 620 /* Note that use of alloca is always safe here. It's obvious for systems | 637 /* Note that use of alloca is always safe here. It's obvious for systems |
| 621 that do not have true vfork or that have true (stack) alloca. | 638 that do not have true vfork or that have true (stack) alloca. |
| 622 If using vfork and C_ALLOCA it is safe because that changes | 639 If using vfork and C_ALLOCA it is safe because that changes |
| 623 the superior's static variables as if the superior had done alloca | 640 the superior's static variables as if the superior had done alloca |