In your example, the term 'would' is used as part of counterfactual. What this means is that one cannot use the normal conditional "if it IS more than a hundred" because it contradicts the assertion there is a limit. From Wikipedia:
Examples
The difference between indicative and counterfactual conditionals can be illustrated by the following contrast:
Indicative: If it is raining right now, then Sally is inside.
Counterfactual: If it were raining right now, then Sally would be inside.
The verb mood when used like this is the subjunctive. In fact, your example counterfactual conditional uses the pluperfect subjunctive to be more specific. From Wikipedia's article on subjunctive:
The English subjunctive also occurs in counterfactual dependent clauses, using a form of the verb that in the indicative would indicate a time of action prior to the one implied by the subjunctive. It is called the past subjunctive when referring counterfactually to the present, and is called the pluperfect subjunctive when referring counterfactually to the past.
As to why is this construction used, instead of the indicative? A fluent English speakers conveys the fact that the line can't actually be more than 100 characters entered in the software, but at the same describes it correctly as it is 100 lines outside the software. Thus, it communicates the possible states of the line both inside and outside, and tackles the assumption that it could be more than 100 by grammatically indicating it is a counterfactual. More simply, the 'would' be 100 implies it 'can't' be 100.
Could the passage be written without using a conditional counterfactual and using the subjunctive? Yes. However, it impoverishes a language to remove more complicated constructions. Consider 'can' and 'may'. More speakers than not use 'can' to ask questions of permission. But, then speakers have to rely on context to determine whether the use of can indicates permission or ability. If one uses may, there is certainty. When one asks 'May I go to the restroom?', one sees the request for permission (and avoids the usual silliness of retorting "I don't know, can you?" to the question "Can I go to the restroom?")
See also the article on the past subjunctive.
EDIT TO REBUT CRITICISM:
LOGICAL BASIS FOR COUNTERFACTUALLITY
"Code, in the reality of the real world, has no line limit." Ontologically speaking, reality is a rather disputatious subject; however, this statement would be rejected by most because it would entail the physical construction of a Universal Turing Machine. Strictly speaking, it is not possible because there are no infinite resources for the construction of any machine, a requirement of the TM.
In fact, the majority of computer scientists I know reject actual infinity for the potential sort. A major point of the computer science curriculum is to come to grips with temporal-spatial constraints of a computer. There aren't actually an infinity of numbers, but rather a number of any size can be potentially generated.
As for wrapping a line, that is essence splitting a string in some sense. Line wrapping in an editor takes a string and prints the string as multiple strings on the display. The source and destination strings actually exist in different buffers, and are physically speaking, different strings with different addresses. Not to get too technical on formal languages, but if the input exceeds the limit of procedure that transfers the string to the application's stack (from wherever to wherever in the architecture), what's really happening is that the original string is programmatically copied into multiple strings using a metalanguage. Thus, one has application-level strings and file-system strings in wrapping. For example, the former might be SWIFT's formatting of the input: "P1_Input%EOL%P2_Input%EOL%P3_Input%EOL%\n" from the input "P1_InputP2_InputP3_Input\n"...
If Len("P1_InputP2_InputP3_Input\n") > STRING_MAX then to have in the application is counterfactual. It would be allowed, if one were to consider "P1_Input%EOL%P2_Input%EOL%P3_Input%EOL%\n" an equivalent string, but strictly speaking it is not. It's not uncommon to think of computation as essentially unlimited, but a careful study of the subject will reject that.
Besides, the author clearly uses the subjunctive. The subjunctive is the biggest clue to the nature of the illuctionary intent of the author. It is certainly not unreasonable on a small project documented by a coder that such a distinction between input and output strings would drawn.
