Gcc the problems in solaris I have a third party because the 1986-1991 Judge CC compiler is used solaris (Sun OS 5.8), while some applications using gcc 3.3.2 compiler, there's a problem : double alignment, gcc default is 8 bytes of double data alignment. CC and the four byte alignment. My question is how to install gcc parameter to enable them to double to 4bytes way alignment. Thank you -no-align-double Not only support intel? In solaris on how no choice? How would you deal with it? -mno-unaligned-doubles -munaligned-doubles Assume that doubles have 8 byte alignment. This is the default. With -munaligned-doubles, GCC assumes that doubles have 8 byte alignment only if they are contained in another type. or if they have an absolute address. Otherwise, it assumes they have 4 byte alignment. Specifying this option avoids some rare compatibility problems with code generated by other compilers. It is not the default because it results in a performance loss. especially for floating point code. I do not know if this is not your needs. --> Thank you, I now need to know how to use this switch in solaris. --> In order not on the trip? --> Checked that no solaris on the gcc compiler switches. --> Your gcc version? I was given that the latest 4.1.0. There, just these two options? Gcc 3.3.2 Can you confirm that the Sun OS 5.8 on these two options? --> These two options and OS versions, but of the gcc compiler for SPARC machines. I have no manual gcc 3.3.2, it is impossible not to know there are two. But I should have guessed. There can be used to test know, the test is it so difficult? I do not understand the question you asked to know what think. If not, I would explain to you not apply. Em05 : : --> I tested that at any time can always easy to prove, not total proved difficult. Besides these two switches Why not available in solaris is the root of the problem. Q I understand now what? Some will simply say it is clear that the sun does not provide long gcc against double/long double/long 4 byte alignment, it would be desirable to talk about why. --> Then you prove what is the result? Now, I did not see that I was given two options can be used or not used. I was given two options from here today. Http://gcc.gnu.org/onlinedocs/ I can only say that these two switches do not provide gcc 3.3.2. As for the versions, or simply impossible to provide, I do not have an answer to that. Gnu you look at the document carefully, there are circumstances in the alignment of the sun, presents three solutions, but not the reality. --> Let me give detailed look at the address. Check tomorrow to work out, do not --> Http://gcc.gnu.org/onlinedocs/gcc-4.1.0/gcc/SPARC-Options.html#SPARC-Options This is the option supported by gcc 4.1.0 on SPARC machines. There were the -mno-unaligned-doubles -munaligned-doubles Describe what you mean is that these two options put double4 byte alignment requirement is achieved? I had misunderstood it is a description? You say I am a bit confused. --> Not understand? 4 gcc who simply try on the sun that told me no problem, can be used on it. Who can or can not explain the reasons why. Convenience sake. --> On, and it would not be a long time machines that way? And you found communication difficult. Em15 : : I have machines, but not all versions of the gcc test again. In addition, the two switches you to solve the problem : For example : Solaris own compiler CC as a function of the parameters of the double data type is 4 byte alignment, which is 8 byte alignment gcc (no-align-double may solve the problem). Of course, there are other pairs, and malloc, I found the information (in other os no-align-double use of the recommendation, but do not support this thing in solaris. What wisdom? --> Sorry, if not two, I have no proposals. : Oops : [ Closed window ]