From: <Joe...@pd...> - 2006-08-28 11:02:07
|
>> I have a large project that is built as a set of static libraries. >> At the very end of the build, all the libraries are linked with >> the main module to construct the final executable. This last step >> takes very large amount of time. Is it possible to speed-up that >> last stage? Would turning all the static libs into dlls, be better? >> I am running gcc 3.4.2 and ld 2.15.91. > >Newer version of binutils (which ls is part of) do provide better >algorithms. Check the archive for details. I'd suggest using >binutils-2.16.91 or even 2.17.50 (both of which I have used for >production code for a really long time now). Woah hoah, well that was a nice tip. 2.16.91 didn't change anything, but with 2.17.50 link time goes from >3 = min to <10 sec. Thanks, Joerg |
From: Ross R. <rr...@cs...> - 2006-08-30 20:05:24
|
Earnie Boyd writes: >Do not SPAM the MinGW users privately either. If I hear that you do I >will ask SF for help in stopping you It's one thing if you and Keith Marshall want to prohibit any mention of alternatives to MinGW on these lists, but someone sending private e-mail about such alternatives is not spam and is quite frankly none of your business. Ross Ridge |
From: Earnie B. <ea...@us...> - 2006-08-31 01:22:13
|
Quoting Ross Ridge <rr...@cs...>: > Earnie Boyd writes: >> Do not SPAM the MinGW users privately either. If I hear that you do I >> will ask SF for help in stopping you > > It's one thing if you and Keith Marshall want to prohibit any mention > of alternatives to MinGW on these lists, but someone sending private > e-mail about such alternatives is not spam and is quite frankly none of > your business. > So you are condoning the use of your email address for commercially supported or non-commercially alternatives to MinGW binaries? You want every one tauting a nice product to contact you? Unsolicited private email is SPAM AFAIC. If someone complains to me about being privately contacted from the email list about a product that isn't MinGW I must take action. Earnie Boyd http://shop.siebunlimited.com |
From: Ross R. <rr...@cs...> - 2006-08-31 02:58:26
|
Earnie Boyd writes: > Do not SPAM the MinGW users privately either. If I hear that you do I > will ask SF for help in stopping you Ross Ridge > It's one thing if you and Keith Marshall want to prohibit any mention > of alternatives to MinGW on these lists, but someone sending private > e-mail about such alternatives is not spam and is quite frankly none of > your business. Earnie Boyd writes: > So you are condoning the use of your email address for commercially > supported or non-commercially alternatives to MinGW binaries? You want > every one tauting a nice product to contact you? Unsolicited private > email is SPAM AFAIC. If someone asks for help on a mailing list and someone else attempts to provide help by responding privately by e-mail that e-mail is neither unsolicited nor spam. > If someone complains to me about being privately contacted from the > email list about a product that isn't MinGW I must take action. No one is going to complain to you about receiving a private e-mail like the one Marek sent this list. Users who need help with MinGW only want solutions to their problems, they aren't going to feel harassed simply because someone suggested using something else. They don't need anyone to defend them from well meaning people like Marek offering help that only a MinGW partisan would find offensive. Ross Ridge Ross Ridge |
From: Keith M. <kei...@to...> - 2006-09-01 09:48:40
|
Ross Ridge wrote: > It's one thing if you and Keith Marshall want to prohibit any mention > of alternatives to MinGW on these lists... I don't object to members suggesting alternatives to MinGW tools; nor I did I indend to imply that we should prohibit the posting of such suggestions on these lists. What I *do* object to, is the blatantly misleading manner in which this particular suggestion, of an alternative to MinGW's `ld', advertised it as `a drop-in replacement' for the standard offering; it doesn't support the full gamut of `ld' switches, so it most definitely does not fulfil the implied promise. None the less, it may still be of interest to some users, on the basis of its claimed speed advantage. I think it is good that users are made aware of such alternatives, but please, be honest about their limitations. Regards, Keith. |