Ususally, covalent bond is stronger than ionic bond.  The important measures for the bond strength are melting point and thermal expansion coefficient. Covalent compounds have higher melting points and low thermal expansion coefficient.

Typical example is diamond, which appear to be  perfectly covalent bonding, and it  has  very high heat of fusion,  and consequent very high melting temperature.

Ionic bond is considered as columbic interactions or non-covalent interactions exist between a cation and an anion. Bond strength order: Covalent bond (40 Kcal/mol)>Ionic bond (5-8 Kcal/mol)>Hydrogen bond (1-5 Kcal/mol).

Generally, covalent bond is stronger than ionic bond. The smaller band gap of GaAs may be a result of other facters.

All the indications shows that the covalent bound, which strongly directional is the strongest bond one can have. Those crystals having bond together by by covalent bonds without exception demonstrate very high enthalpies of fusion and sublimation; very high melting points,  high elastic modulus and rigidity etc. Even in the liquid state that they keep their short range  atomic order' 

It will be useful to know that covalent (polar) is due to electronegativity between the two types of atoms atoms while in ionic bonding the same is between a metallic atom and a nonmetallic atom. Without further discussion one can understand that there is no meaning in comparing the two compounds GaAs  and NaCl. I. e,  their physical properties characterization the differences.

Prof S. Devanarayanan 

We first need to analyse how to measure the strength of a chemical bond. Two among the possible measurements will be the band-gap and the melting point. Both band-gap and melting point represent the breaking of the chemical bond. 

Let us consider the melting point data first. Among elements, tungsten has the highest melting point of 3687 K. Carbon (Diamond) does not melt at ambient pressure but sublimes at 4000 K. Diamond is a covalent material, but other covalent elements - Si and Ge - have much lower melting points than the refractory metals - molybdenum, tantalum, tungsten. Based on these data, one cannot say that the covalent elements have the strongest bonds or have stronger bonds than the metallic elements (metallic bonds). In fact, diamond is the only exception among covalent solids. All other covalent materials (elemental as well as compound) have much lower melting point than many ionic or metal oxides and certainly much lower melting points than the refractory oxides - TiO2 (1870 C), ZrO2 (2727 C), MgO (2795 C), BeO (2495 C); these oxides are ionic solids. The highest melting point has the compound tantalum hafnium carbide of 4488 K; this compound has a complex mixture of ionic, metallic, and covalent bonds.  

Let us now consider the band-gap. Among the inorganic solids, the highest band-gap has SiO2 of about 9 eV, but its melting point is 1726 C, lower than the above refractory oxides, which have much lower ban-gaps than SiO2. SiO2 is more covalent than ionic. 

Basically, all chemical bonds are electrostatic in nature. The difference between the ionic and the covalent bond is that in the ionic bond, the cation and the anion are fixed, whereas in the covalent bond cations and anions interchange every instant. Diamond is purely covalent, but, no compound can be purely covalent because of the electronegativity difference between the cation (Ga in GaAs) and the anion (As in GaAs). Hence, all compound semiconductors are partly ionic, and the percentage of iconicity has been observed to increase the band-gap. 

The electronegativity difference DEN is often used to classify the type of chemical bond between atoms. The following is an illustration of this kind of an empirical scheme; the range indicated can be considered only an approximate one. Non-polar covalent bond: DEN \ 0.5; Polar covalent bond: DEN = 0.5–1.6; Ionic bond: DEN [ 2.0.

I am at a loss to understand the comment of Prof S. Devanarayanan: "Without further discussion one can understand that there is no meaning in comparing the two compounds GaAs and NaCl. I. e, their physical properties characterization the differences."

There appears to be no one-to-one correspondence between the band-gap and the melting point - the two most visible signatures of the chemical bond strength. 

I agree with Samares that it is very difficult to find strong correlation between the nature of a  bond and its strength so-called its chemical strength. First of all for me,  the terminology such as the chemical strength of a substance, it doesn't mean any thing as long as it is not given its operational definition. How we are going to measure the chemical strength?  The strength is a mechanical concept  and  every human beings can easily perceive it, and can find a measuring stick, which can be accepted universally anywhere any time.  Situation with the chemical strength is not very clear. It seems there is no a single metric  available that every one may  agree on! 

Probably the enthalpy of fusion for solid substances may a candidate for that measure. Similarly cohesive energy,  which is nothing but the energy of sublimation. Which is highest for the transition metals {Fe, Co, Ni , Cu; Ru, Rh, Pd. Ag; Os, Ir, Pt, Au }  and covalently bounded crystals (C, Si4.66, Ge3.55, Sn3.12,Pb2.04).  The highest among those are  (Os 8.12, C7.35, Ir6.89, Ru6.68, Pt5.85, Rh5.77  eV/Atom ..) . 

The discussion given by Samares is good, and it has brought in bonding in various materials. but has it provide an answer to the cases in GaAs and NaCl ? Pauling's has discussed the differences in H-bond where one side is covalent bonding and the other weakly bonded side. Still the mystery for the question of strength of covalent bonding in GaAs and Ionic bonding in NaCl basically involves the nature of the elemental atoms , viz, Ga, Al,  Na, and Cl. with differing (nuclear) Z-values and electronegativitIes. The two compound have different physical conditions for crystal growth !!!

In this particular case it seems difficult to comment but in general ionic bond is stronger than the covalent bond as the lattice energy for ionic crystals is higher than the bond energy  for covalent compounds..