a full buildworld may be required to compile the latest world (2025)

Discussion:

a full buildworld may be required to compile the latest world

John Marino

2013-06-12 09:49:58 UTC

Permalink

I just updated libm. That's the second time in a couple of weeks this
library has been updated (this time it was to add 8 new sorely-needed
long double functions).

Rather than bump the shared lib major version again, I introduced symbol
versioning for libm. This means that libm will permanently be called
libm.so.4 as it is now, but we'll update the versions of the symbols
internally as needed in the future.

The drawback is that another full world build may be needed in order to
build HEAD. It's possible rtld can still match the symbols but I didn't
test this as I had to do a full build anyway because I was behind the
previous libm/libc bump.

I suspect we will add symbol versioning to more libraries -- libc is a
good candidate for the next one because the bump to libc.so.8 was more
painful than it should have been, but this was partially due to
insufficient TLS storage space which has also been addressed. It also
resulted in binaries linking to both libc.so.7 and libc.so.8 which
symbol versioning would have avoided.

John

John Marino

2013-06-12 09:49:58 UTC

Permalink

I just updated libm. That's the second time in a couple of weeks this
library has been updated (this time it was to add 8 new sorely-needed
long double functions).

Rather than bump the shared lib major version again, I introduced symbol
versioning for libm. This means that libm will permanently be called
libm.so.4 as it is now, but we'll update the versions of the symbols
internally as needed in the future.

The drawback is that another full world build may be needed in order to
build HEAD. It's possible rtld can still match the symbols but I didn't
test this as I had to do a full build anyway because I was behind the
previous libm/libc bump.

I suspect we will add symbol versioning to more libraries -- libc is a
good candidate for the next one because the bump to libc.so.8 was more
painful than it should have been, but this was partially due to
insufficient TLS storage space which has also been addressed. It also
resulted in binaries linking to both libc.so.7 and libc.so.8 which
symbol versioning would have avoided.

John

John Marino

2013-06-12 09:49:58 UTC

Permalink

I just updated libm. That's the second time in a couple of weeks this
library has been updated (this time it was to add 8 new sorely-needed
long double functions).

Rather than bump the shared lib major version again, I introduced symbol
versioning for libm. This means that libm will permanently be called
libm.so.4 as it is now, but we'll update the versions of the symbols
internally as needed in the future.

The drawback is that another full world build may be needed in order to
build HEAD. It's possible rtld can still match the symbols but I didn't
test this as I had to do a full build anyway because I was behind the
previous libm/libc bump.

I suspect we will add symbol versioning to more libraries -- libc is a
good candidate for the next one because the bump to libc.so.8 was more
painful than it should have been, but this was partially due to
insufficient TLS storage space which has also been addressed. It also
resulted in binaries linking to both libc.so.7 and libc.so.8 which
symbol versioning would have avoided.

John

John Marino

2013-06-12 09:49:58 UTC

Permalink

I just updated libm. That's the second time in a couple of weeks this
library has been updated (this time it was to add 8 new sorely-needed
long double functions).

Rather than bump the shared lib major version again, I introduced symbol
versioning for libm. This means that libm will permanently be called
libm.so.4 as it is now, but we'll update the versions of the symbols
internally as needed in the future.

The drawback is that another full world build may be needed in order to
build HEAD. It's possible rtld can still match the symbols but I didn't
test this as I had to do a full build anyway because I was behind the
previous libm/libc bump.

I suspect we will add symbol versioning to more libraries -- libc is a
good candidate for the next one because the bump to libc.so.8 was more
painful than it should have been, but this was partially due to
insufficient TLS storage space which has also been addressed. It also
resulted in binaries linking to both libc.so.7 and libc.so.8 which
symbol versioning would have avoided.

John

John Marino

2013-06-12 09:49:58 UTC

Permalink

I just updated libm. That's the second time in a couple of weeks this
library has been updated (this time it was to add 8 new sorely-needed
long double functions).

Rather than bump the shared lib major version again, I introduced symbol
versioning for libm. This means that libm will permanently be called
libm.so.4 as it is now, but we'll update the versions of the symbols
internally as needed in the future.

The drawback is that another full world build may be needed in order to
build HEAD. It's possible rtld can still match the symbols but I didn't
test this as I had to do a full build anyway because I was behind the
previous libm/libc bump.

I suspect we will add symbol versioning to more libraries -- libc is a
good candidate for the next one because the bump to libc.so.8 was more
painful than it should have been, but this was partially due to
insufficient TLS storage space which has also been addressed. It also
resulted in binaries linking to both libc.so.7 and libc.so.8 which
symbol versioning would have avoided.

John

4 Replies
1 View
Permalink to this page
Disable enhanced parsing

Thread Navigation

John Marino2013-06-12 09:49:58 UTC
John Marino2013-06-12 09:49:58 UTC
John Marino2013-06-12 09:49:58 UTC
John Marino2013-06-12 09:49:58 UTC
John Marino2013-06-12 09:49:58 UTC
a full buildworld may be required to compile the latest world (2025)
Top Articles
Latest Posts
Recommended Articles
Article information

Author: Carmelo Roob

Last Updated:

Views: 6417

Rating: 4.4 / 5 (65 voted)

Reviews: 80% of readers found this page helpful

Author information

Name: Carmelo Roob

Birthday: 1995-01-09

Address: Apt. 915 481 Sipes Cliff, New Gonzalobury, CO 80176

Phone: +6773780339780

Job: Sales Executive

Hobby: Gaming, Jogging, Rugby, Video gaming, Handball, Ice skating, Web surfing

Introduction: My name is Carmelo Roob, I am a modern, handsome, delightful, comfortable, attractive, vast, good person who loves writing and wants to share my knowledge and understanding with you.